Tobacco Documents Online

Smoking and Health 1964-1979 T H E C 0 N T I N U I N G C 0 N T R 0 V E R S Y T H E T 0 B A C C 0 I N S T I T U T E 1776 K Street, N.W., Washington, D.C. 20006 January 10, 1979

Table of Contents Page Preface v Overview -- Smoking and Health 1979 1 Public Smoking 13 Smoking and Over-All Mortality 35 Women and Smoking 47 Cancer in the Work Place 71 Lung Cancer 87 Other Cancers 103 Cardiovascular Disease 119 Chronic Obstructive Pulmonary Disease 139 Appendix 153 This volume is published by The Tobacco Institute in the belief that public discussion about tobacco smoking is in the public interest and that the smoking controversy must be resolved by scientific research. 111

Pref ace The American people would be better served if high government health officials and private interest groups which encourage them abandoned the myth of "waging war" against diseases and their alleged causes. The process of making public policy is better served when areas of scientific unknowns are illuminated by the light of reasoned deliberation rather than the heat of emotional rhetoric. Nature will not yield her secrets to media events, propaganda barrages, self-righteous zeal or official fiat. The enigma of cancer and chronic diseases will yield only to the steady advance of scientific knowledge. And know- ledge does not flourish in a lock-step society. It grows best under conditions of unfettered investigation and free, fair and full discussion. Indeed, many scientists are becoming concerned that preoccupation with smoking may be both unfounded and dangerous -- unfounded because evidence on many critical points is conflicting, dangerous because it diverts attention from other suspected hazards. It should be noted that plans for the first report of the Surgeon General's Advisory Committee on Smoking and Health in 1964 called for "the study [to] be concerned not rQ a only with tobacco, but all other factors which may be involved i -p w such as air pollution, automobile exhausts, etc." N -~s V

One does not become an advocate of tobacco by sup- porting a broader, deeper, more objective consideration of the issue. Over a hundred years ago, William Hazlitt, the English critic and essayist, put his finger on the nub of the problem. You may agree with him when he said: The origin of all science is in the desire to know causes; and the origin of all false science and imposture is in the desire to accept false causes rather than none; or, which is the same thing, in the unwillingness to acknowledge our own ignorance. It is time for all parties to this controversy to admit that there is much that is unknown. Doing so will en- courage research to reduce the deficit in our knowledge and In that spirit, we offer for consideration this document, which -- while not intended to be exhaustive -- raises some of the questions in the continuing smoking and health controversy. .p ' January 10, 1979 vi

Overview -- Smoking and Health 1979 Fifteen years have passed since the release of "Smoking and Health" -- the first and perhaps the most widely publicized of a series of such reports prepared by the Depart- ment of Health, Education and Welfare. Despite millions of dollars spent since that time both by the government and the tobacco industry on smoking and health-related research, many questions about the relationship between smoking and disease remain unanswered. Now, as in 1964, there are statistical relationships and several working hypotheses, but no definitive and final answers. Despite claims to the contrary, no one -- in govern- ment or industry -- can explain the reported associations of smoking with lung cancer, heart disease, emphysema, low infant birth weight, and yes, even cancer of the pancreas. No one knows why -- or how -- a cancerous growth begins, whether it is in the lung, pancreas, or bladder. No one knows why the walls of human arteries become clogged with lipids or how clots that can lead to stroke get their start. No one knows why pregnant women who smoke have lighter infants on the average than women who don't smoke, or why some women, whether or not they smoke, have smaller 1

inf ants . Scientists have not proven that cigarette smoke or any of the thousands of its constituents as found in ciga- rette smoke cause human disease. Nor have scientists demonstrated that the healthy nonsmoker is harmed by his neighbor's cigarette smoking. But because some agencies in the U.S. government, members of the medical profession, and others who just don't like cigarette smoke act and react as if all the claims about smoking are sci_entific certainties, The Tobacco Institute sets forth here certain evidence which relates to such judg- ments. A comment made by the U.S. Surgeon General in his foreword to the 1964 report is as relevant today as it was 15 years ago: The interrelationships of smoking and health undoubt- edly are complex. The subject does not lend itself to easy answers. Nevertheless, it has been increas- ingly apparent that answers must be found. cn 0 ~ Public Smoking ~' ~ w Other people's smoke has never been shown to cause r~~.~ disease in nonsmokers. ~ ~ Scientists, researchers, government officials and even some well-known anti-smoking spokespersons have stated 2

that smoking in public places does not harm the healthy non- smoker. Some persons may find the tobacco smoke of others annoying in some circumstances. Last year, a study conducted by Danish and British researchers found "transitory discomfort" but no evidence of lasting adverse health effects rette smoke in otherwise healthy individuals. from ciga- Some persons who favor banning tobacco smoke in public places cite an article published last year claiming that exposure to cigarette smoke resulted in changes in the exercise perfor mance ability of patients with severe angina pectoris. What is usually ignored is that this study is subject to severe criticism for faulty design as well as unsupported conclusions based on patients' self-described symptoms. Some nonsmokers claim to be allergic to tobacco smoke. However, neither cigarette smoke nor any of the com- ponents as found in cigarette smoke has been demonstrated to be a human allergen. Nonsmokers who make such a claim some- times cite a study which shows that smokers as well as non- smokers react positively to skin tests with tobacco leaf extract, but this is an inappropriate substance to use in allergy testing for tobacco smoke. Another claim frequently made by anti-smokers, that children are harmed by their parents' smoking, is mainly based on several studies published in the late 1960s and 1970s suggesting that cigarette smoke may be responsible for 3

adverse effects in children. However, questions have been raised about both the experimental methods and the reliability of the conclusions. Moreover, a number of recent studies have failed to demonstrate adverse effects in children of smoking parents. Other people's smoke has never been shown to cause disease in nonsmokers. Over-All Mortality The use of results from flawed population studies to frighten people by attributing large numbers of deaths yearly to smoking may be misleading and is most regrettable. ~ Assertions that nonsmokers as a group live longer than smokers are based on studies that were poorly designed and statistically flawed. For example, they involved samples not representative of the general U.S. population. Despite these problems, data from the reports are still used to sup- port a variety of claims about smokers' mortality, including the charge that several hundred thousand Americans die each year because they smoke. With such use -- and misuse -- of data, it is probably not surprising that a caveat in the 1964 Surgeon General's report is often overlooked: "Statistical methods cannot establish a causal relationship..." ~ 0 ~ 4 N 10

Women and Smoking Inconsistent findings from studies of smoking women and their children make it impossible to draw con- vincing conclusions from the data. Pregnancy Outcome Although the abbreviated 1977-78 HEW report to Con- gress concludes that cigarette smoking is "probably causally associated" with increased perinatal mortality, it relies on data which indicate that any claims of a causal relationship have a highly questionable foundation. The data suggest that such factors as history of previous pregnancy loss and hospital pay status (public vs. private) have greater effects on preg- nancy outcome than maternal smoking. The data provide support for the belief that adverse pregnancy outcomes -- indeed, the health and life or death of the child itself -- may be pre- determined by who the mother is -- her constitution or innate characteristics -- rather than whether or not she smokes. Smoking and Early Menopause Research which appears to indicate that smokers undergo menopause earlier than nonsmokers has been used to support a claim that smokers are depriving themselves of the "protection" from heart attacks believed to be provided by female sex hormones until change of life. 5

However, this claim is not supported by heart disease mortality statistics which show no "jump" during the menopausal age span -- such as might be expected if large numbers of women were suddenly deprived of "protection" against this disease. The almost single-minded concentration on smoking evident in much research in this area may result in a failure to consider other factors that may be involved. For example, Public Health Service research indicates that menopause begins earlier in black women, in white women from lower income levels and rural areas and in leaner women. Oral Contraceptives The scientific literature does not support the claim that oral contraceptive users who stop smoking decrease their disease rates significantly. This point was recently made in a Congressional hearing during which a decision by the Food and Drug Administration to require a printed warning -- which, in. effect, implies such an assurance -- came under attack. In discussions on this issue, concern has been expressed that the reported statistical relationship between oral contraceptive use, smoking and illness may cause scien- tists to overlook other factors that may explain this relation- ship. 2501443131 6

Women and Lung Cancer A comparison of international lung cancer patterns raises serious questions about the claim that the larger number of women smoking today accounts for their rising lung cancer rates. Such a comparison indicates,that lung cancer patterns in women are different in various countries and are dissimilar from male patterns. Even when allowance is made for the later popularity of smoking among women, there is no consistent trend of increasing mortality rates. Some scientists have questioned whether the recent increase in lung cancer mortality is more artificial than real. These queries have been made because physicians appear to be ordering more diagnostic tests for women they know to be smokers. Therefore, it may be possible that more lung cancer among women is being diagnosed because more reliance is being placed on diagnostic techniques not available in the past. In addition, a role for occupational and/or other environmental exposures has been suggested by research con- ducted in heavily industrialized counties. Cancer in the Work Place The almost exclusive focus on individual smoking habits in the study of disease may have delayed needed research into possible occupational and environmental causes. tV Cn 0 ~ ~ ~ w N N 7

The announcement last September by HEW Secretary Califano that at least 20 percent of all cancers may be occupa- tionally related brought angry denials from anti-smoking researchers and organizations whose own estimates differed significantly from the new estimate. The authors of the report referred to by the Secretary actually estimated that between 21 and 38 percent of all cancers were occupationally related. They attributed a sizable proportion of all occupational cancers to asbestos exposure and noted that "perhaps the most important lesson to be learned from the asbestos story is that a major health disaster can develop while its early manifesta- tions are lost by being attributed to other factors." Lung Cancer The failure to consider critically 1) important diag.nostic advances, 2) cha_nges in the r_eQorted frequencies of lunK cancer cell types and 3) trends in cigarette consumption and lunJg cancer mortality data raises serious questions about any conclusions regardinR smoking. What some have called the "epidemic" in lung cancer mortality in this century has been linked by some to the increased popularity of smoking. However, it has been specu- lated that this reported increase may in fact have been created largely by improvements in diagnostic techniques -- in other words, more lung cancer cases have been reported because 8

physicians were better equipped to find them. Even if at least a portion of the "epidemic" is real, trends in lung cancer death rates can not be satisfactorily explained by cigarette consumption patterns. A British researcher has reported a similar phenom- enon in the United Kingdom that he contends is more consistent with the constitutional hypothesis than the smoking-causation theory. An apparent change in the reported frequencies of lung cancer cell types was observed in 1977. A researcher noted a shift in the histologic types of lung cancer found at a major cancer center -- from epidermoid, or squamous-cell, which has been more strongly associated with cigarette smoking, to adenocarcinoma, which is only weakly associated. Up to this point, at least, adenocarcinoma has been the predominant hist'ologic type of lung cancer reported in women and in non- smokers. Because of this reported increase in the lung cancer cell type not generally associated with cigarette smoking, serious doubt is cast on the role of smoking in the development of this disease. Other Cancers The establishment of an Y relationship between smok- ing ng and cancers of the larynx, esophagus and bladder must involve considerable guesswork, because of the vastly different incidence patterns and trends of these diseases and multiQle susQected causes.

Incongruities in the trends of incidence rates for "other cancers," such as cancers of the oral cavity, bladder, esophagus, larynx and pancreas, are almost impossible to reconcile with the cigarette smoking causal hypothesis. These trends, as described last year, include such anomalies as incidence rates that rise, fall or remain stable depending on disease, sex and race. Moreover, new evidence indicates that a number of these cancers may be associated with alcohol consumption and that the association of alcohol with cigarette smoking may not only confound the relationships but may hide other correla- tions. In addition, recent work has implicated occupational exposures and nutritional factors in the etiologies of some of these cancers. Cardiovascular Disease tv ~ 0 ~ A fair appraisal of the evidence, examined in its -~P , entiretY, indicates that the risk of coronary heart disease is strongly associated with genetic CO Cn and lifestyle factors. In 1977, the director of the governmental agency responsible for cardiovascular research told a Congressional committee that "we still don't know the etiology of arterio- sclerosis and hypertension" and that his researchers are still testing the "hypothesis...that lowering cholesterol and cessation of smoking will delay or prevent the onset of 10

heart disease" (emphasis added). Meanwhile, statisticians were finding that death rates for heart disease continued the decline that began in the late 1960s -- in all age groups, in both sexes and in both whites and nonwhites. There was new evidence in 1976, 1977 and 1978 that lifestyles, personality patterns and hormonal imbalances are implicated in coronary disease. An important development in cardiovascular research, was reported in 1977, by a group of researchers who reported that they were unable to duplicate their previous findings, which they said had suggested a causal role of carbon monoxide (CO) in the development of cardiovascular disease. In a presentation describing their findings, they said "no direct toxic effect of CO" could be demonstrated. Chronic Obstructive Pulmonary Disease The uncertainties and unknowns in the medical under- standing of COPD permit no firm conclusions about smoking . Chronic bronchitis and emphysema are highly complex and poorly understood diseases. Despite serious gaps in the medical knowledge in this area, claims abound that these diseases are caused by smoking. The validity of such claims is challenged by a recent National Heart, Lung and Blood Insti- tute statement that "the exact etiology of emphysema and other chronic lung diseases is unknown..." 11

Public Smoking This could have been the shortest chapter in this collection, considering the first item to be presented below. But it won't be, because the myths which have grown up around the whole subject of public smoking (sometimes called "passive smoking") must be discussed. Despite those myths, the recent testimony of the man who has directed the government's smoking and health research program for more than 10 years may well summarize the situation. Dr. Gio B. Gori of the National Cancer Institute was asked by a Congressional committee last October, "Is Other people's smoke has never been shown to cause disease in nonsmokers. there evidence to suggest that there may be an increase in risk of heart and lung disease to a nonsmoker by being in the presence of smokers?" Replied Gori: Well, this is a difficult question to answer, Mr. Chairman, because the answer that I have to give as a scientist may not please always the anti-smoking forces. But the fact remains that we really do not have conclusive scientific evidence about the adverse health effects of passive smoking on the bystander [and] in the usual conditions under which smoking is practiced, the evidence does not indicate that the casual bystander is seriously harmed by smoking (1). 134,

Very recent research -- two papers published in the second half of 1978 -- confirms this appraisal. In one, which appeared in the American Medical Association's Archives of Environmental Health, a Canadian team found that physiological responses to smoke exposure in normal adults could be described only as "minimal" and said "arguments concerning effects rest on symptomatology" -- in other words, on how the subjects said they felt, a highly unreliable standard (2). In the second, Danish and British scientists wrote in International Archives of Occupational and Environmental Health that they found "transitory discomfort" but that con- stituents of neither the gas phase nor the particulate phase of ambient cigarette smoke have "a lasting adverse health effect in otherwise healthy individuals" (3). Like the Canadians, the researchers said their test situations were realistic and typical of rooms in which smokers gather. Carcinogenic effects have not been demonstrated in humans. Nor do any studies to date establish that breathing others' tobacco smoke either causes lung disease or worsens the status of patients with existing disease (4). And it has not been established that atmospheric tobacco smoke has a causal role in coronary heart disease in nonsmokers (5). What evi- dence there is that ambient cigarette smoke may affect CHD sufferers has been called deficient (see below). 14

Nitrosamines A recent technique of some researchers has been to measure levels of specific cigarette smoke components in various public places and then announce that the non- smokers, by merely being present, would inhale of so many cigarettes in so many hours. the equivalent One such experiment in New York was described in 1977 at a joint conference of the American Chemical Society and the Chemical Institute of Canada (6). One of the researchers told how they had isolated tiny amounts of nitrosamines in cigarette smoke from laboratory smoking machines. Then they designed an apparatus to measure the compounds also in the smoke produced between the machine's puffs (the sidesteam smoke, in effect). Next they fitted the apparatus, consisting of two glass jars full of "trapping solution", some tubing and a battery-operated vacuum pump, into an attache case. And they headed, attache case in hand, first for a smoky New York commuter train bar car and then for a small metropolitan bar "frequented primarily by cigarette-smoking clientele." The amounts of nitrosamines they trapped in their glass bottles, they told their fellow chemists, indicated that customers in the bar could have inhaled in an hour the

same amount of nitrosamines as a smoker of nine to 16 nonfilter cigarettes. The researchers then discussed the "shortcomings" of their experiment. First of all, they were dealing with "volatiles" which change swiftly in the atmosphere (if not in a glass bottle with "trapping" solution). Then there were the two assumptions on which their estimates were based: one, that the mode of inhalation in breathing and smoking are comparable and, two, that "man's smoking conditions are synonymous with those of the test machine." Then they admitted that neither assumption was true. And they have yet to publish their results in a scientific journal. The measurement of nitrosamines in tobacco smoke has been controversial. Even the 1971 HEW report to Congress commented that nitrosamines reportedly found in smoke may be "artifacts dependent on the method of smoke collection" (7). A University of California chemist has estimated that at the trace amounts nitrosamines were reported in side- stream smoke, smoking as many as 100 cigarettes in a small room would produce only 3 nanograms (3 one-billionths of a gram) of the compound per liter of air (8). He said that at this concentration a carcinogenic effect could not be demon- PO strated in animals. Cn O ~ A 41 W ~ 1.~' Q

Carbon Monoxide Th ere was considerable furor last July at the publi- cation of an experiment by an avowedly anti-cigarette resear- cher in California -- coincidentally just before that state voted on a proposed statewide restrictive public smoking measure. Dr. Wilbert S. Aronow wrote that remaining two hours at a time in smoke-laden rooms adversely affected 10 patients with angina pectoris (9). The research has been roundly criticized since then for faulty study design as well as unsupported conclusions based on self-described symptoms of severely ill patients who had no doubt been advised, as are most angina sufferers, to avoid stressful situations (5, 10-16). Calling the study "flawed," a Los Angeles chest specialist wrote in The Los Angeles Times: Both investigators and subjects were aware of their exposure to the smoke and, obviously, that such exposure might be harmful. Moreover, the major measurement of the study -- the occurrence of chest pain -- was subjective. In other words, the test subjects' reports of chest pains while exercising could well have been influenced by their belief that they had been harmed by exposure to cigarette smoke. Considering this, as well as the curiously low variation in the subjects' carboxyhemoglobin levels [with and without exposure], I find the study questionable and sorely in need of confirmation (17). N Cn 0 ~ -p ~ W ~ ~ ~ 17

Carboxyhemoglobin (COHb), referred to by the Los Angeles physician, is the key to the charges made against cigarettes in both the heart disease area and in public smok- ing. COHb is created in the blood when carbon monoxide, from whatever the source, combines with red blood pigment. Claims about the alleged health effects of CO as found in cigarette smoke (see also Cardiovascular Disease) appear to be based on research which shows that an average smoker has higher levels of COHb than an average nonsmoker. Experiments in test rooms under varying atmospheric conditions, with varying numbers of cigarettes and other tobacco products burned or smoked by volunteers, have shown increases of CO in the room air. An important point to re- member is that when indoor areas are adequately ventilated, CO levels should be reduced to a point that no adverse effects will occur in persons and groups expected to be present in such environments (18). The few studies that also measured the COHb levels in smokers and nonsmokers in these test rooms found that even under severe conditions there was not an appreciable increase in COHb (19). In what might be considered the most extreme experi- ment, incidentally, four persons got into a small European car, windows up tight, inside a closed garage (20). Two =t%xj 18

smokers smoked 10 cigarettes rose sharply in all subjects. in an hour, and the COHb levels Studies have measured CO in the atmosphere under more realistic conditions. Combined results of these studies indicate that CO from smoke measured in normal daily situations rarely exceeds 10 parts per million (21-26). Cigarette smoke is, of course, not the only source of CO in the atmosphere. The most predominant modern source is exhaust from the internal-combustion engine (27). CO is also a natural body constituent created by normal metabolism. Recently a Canadian researcher reported that high levels of CO are generated when pots and pans are placed over otherwise clean burning gas flames during the cooking of a meal (28). Nicotine The question of ni*cotine absorption by nonsmokers has been investigated by some scientists. These studies, in both experimental and real-life situations, have shown that nonsmokers are exposed to insignificant amounts of tobacco smoke. In fact, when a German researcher found that the nonsmoker takes up only a small fraction of nicotine, he concluded that "when speculating on possible health hazards by passive smoking, one may ignore nicotine" (29). N cn 0 ~ In order to discover the possible physiological Pb 19

effects of exposure to nicotine, other researchers recorded heart rate, blood pressure, electrocardiogram readings and skin temperature in nonsmokers in experimental rooms with heavy concentrations of tobacco smoke. None of these para- meters was measurably affected (30). Work Performance Any claims that tobacco smoke, or the CO in tobacco smoke, adversely affects work performance are not warranted by the evidence. The HEW 1975 report to Congress said that what evidence there is in that area is conflicting and any "psycho- motor" effect of CO exposure "remains unclear" (19). The Federal Aviation Administration addressed the point recently in its denial of a petition that smoking be banned in the cockpit (31). In its denial, the FAA cited conclusions of Air Force scientists on effects of abnormally high levels of CO on performance. The agency's news release about the petition denial said: FAA said the information submitted by the petitioners to support their contention that smoking impairs performance "is too inconclusive to warrant the issuance of the requested rule"...FAA conceded that smoking can reduce the blood's oxygen-carrying capability but said there is no evidence at present that this has any deleterious effect on performance. In fact, the agency noted that there is evidence that the body adapts to the effects of small amounts of carbon monoxide by increasing its red blood cell 20

mass and, thus, its oxygen-carrying capability. It called the petitioners' failure to address this point a major deficiency in their argument (32). A National Institutes of Health panel of scientists also considered the question of cockpit smoking early last year at the request of the Surgeon General, perhaps as a result of the FAA turndown and the FAA's view that the original petition was deficient. After a thorough review of the scientific literature as well as information obtained from the Department of Defense and the FAA, the NIH experts came to the same conclusion as the FAA. Smoking by the pilot -- even if he or she is a regular smoker -- they said, "is judged to have negligible effects on physiological and psychomotor functions and there- fore on flight safety, especially when compared to overall performance skills demanded of pilots" (33). Smoking in airplanes brings up another aspect. Is it the cigarette smoke itself or the sig_ht of cigarette smoke that bothers the typical person who objects to public. smoking? Recently, a trial plan in which smokers sat across the aisle from nonsmokers was tried by Western Airlines. ~ ~ Western engineers said the trial produced a torrent of com- o f~t A plaints and blamed psychological effects rather than any ~ N 4h- ` cn ~21

ventilation problem (34). The former aviation editor of United Press International, commenting on that Western trial run, said, "It seems that all a nonsmoker needed was to see someone smoking, and that was enough to make him think he could smell the smoke. I'm afraid this is just one more instance where emotionalism gets in the way of established scientific facts" (35). Allergy? Some nonsmokers, of course, say they are not just bothered by smoke, they are also allergic to it. Indeed, former Surgeon General Luther Terry stated in 1977 that "there are a few people who are actually allergic to tobacco smoke and can be made ill from exposure" (36). Despite Terry's belief, neither cigarette smoke, nor any of its components as found in smoke, has been demonstrated to be a human allergen. Researchers at the Mayo Clinic reported to the American Academy of Allergy in 1976 that they had tried, and failed, to find evidence of tobacco smoke allergy in their tests with patients (37). Last year, after analyzing the pertinent literature and discussing his own research work at r1) ch Tulane Medical Center, Dr. John Salvaggio told a Congressional ~ .~ committee that "there is no proof that tobacco smoke is aller- ti ~ ~ . m 22

genic in man" (38). Claims about tobacco allergy stem primarily from research that has been done with tobacco leaf. Extracts of leaf produce allergic reaction in some people -- both smokers and nonsmokers -- usually those who are otherwise allergic (39-42). Whether cigarette smoke or any of its constituents is allergenic cannot be resolved, however, in tests with tobacco leaf. Dr. Carl Becker and colleagues from Cornell Medical College have reported isolating a brown pigmented molecule (termed "tobacco glycoprotein") from both cured leaf and from smoke. They claim it is an allergen (43-45). These experi- mental pathologists have speculated that "tobacco glycoprotein" may be responsible for some of the diseases which have been statistically associated with smoking. Becker's reports were of special interest to Russell Stedman, who for years directed work with tobacco leaf and tobacco smoke condensate at the Department of Agriculture Eastern Regional Research Center in Philadelphia. Indeed, Stedman, who is retired now and serving as a biochemical consultant to Temple University, tried to duplicate the published isolation procedures of Becker. Stedman said he P\) found that one step in Becker's method introduced as yet cn 0 unidentified extraneous contaminants into the brown pigmented ~ co N 23

material isolated, then tested by Becker for allergenic proper- ties. And, Stedman wrote last year to the House Agricul- ture Tobacco Subcommittee (46) that Becker's separation technique had already been documented, as early as 1971, as introducing "substantial chemical materials" into what was tested (47). In his statement, Stedman contended that for a number of reasons he was "unconvinced" that Becker had demon- strated the presence of a human allergen in either tobacco smoke or smoke condensate. For example, Stedman wrote that Becker has not specified whether the "glycoprotein" with which he had produced allergic reactions in volunteers was from tobacco leaf or tobacco smoke condensate. Although Becker assumed that the materials were identical, Stedman noted that he had concluded from his own research that they were quite different chemically. Stedman also stated that he was not sure whether the material Becker had extracted even was a glyco- protein. Parent/Child Effects A favorite slogan recently of those urging legisla- tion to restrict public smoking has been that cigarette smoke can cause respiratory disease in a child. Or if they want to be more dramatic, anti-public-smokers say something catchy, such as the declaration of a New York., physisisp that „°Th_e 24

largest area of child abuse is parental smoking" (48). That smokers' children have more respiratory ailments because their parents smoke is strictly conjectural. Much epidemiological research has been done on the question here and abroad, but no investigator has been able to demonstrate that cigarette smoke in the home is responsible for a child's picking up a germ or cough. Some have reported that cigarette smoke may be responsible for adverse effects in youngsters (49-53). But their work has been questioned because of faulty study design or suspect conclusions (19, 54). Epidemiologists whose own research in smoking and nonsmoking homes in three American cities was published in 1977 (55) wrote on the subject again in 1978 in British Medical Journal (56). First they summarized their earlier conclusions that neither lung function nor respiratory symptoms of a nonsmoking husband or wife were affected by a smoking spouse and that parental smoking appeared to have "no effect on children's respiratory symptoms" or "lung function." Then they suggested that "the only definite evidence" that parental smoking may affect children's respiratory systems was a British study (57) indicating that the infants of smoking parents have more respiratory illness during the N cn O ~ first year of life than infants of nonsmoking parents. -~a ~ W ~ -p NO 25

But they pointed out that in the same population this was not true in children aged one to five years. And they concluded that "at present there is no firm evidence that these illnesses in children under one year have a serious and lasting effect as no excess of respiratory illness or diminished lung function has been found in the older children of smoking parents." At least four other studies have failed to demon- strate adverse effects on children of smoking parents (58-61). Other environmental and socioeconomic factors that have been associated with respiratory symptoms and diseases include a "cooking effect" identified in a four-year study of 5,700 youngsters in England and Scotland (62). Boys and girls from homes in which gas was used for cooking had more coughs, "colds going into the chest" and bronchitis than children whose homes had electric stoves. The researchers concluded that products of fuel combustion might be the cause of the increased respiratory illness. Conclusion Public smoking has not been shown to cause disease in nonsmokers. As a past president of the American Association for Thoracic Surgery recently stated: An assertion that tobacco smoke is a health hazard 2b

to the normal nonsmoker is untenable. The weight of evidence as it exists in the world literature does not support a claim of adverse health effects for those exposed to "passive smoking" (63). Speculation that reaction to public smoking may have psychological or emotional origins is worth considering. A medical columnist noted recently that "symptoms may come from anger rather than from the smoke itself" (64). He added that: ...what is irritating or annoying has not yet in any scientific study been shown to be dangerous to the nonsmoker. N tn 0 ~ -p ~ ~ ~ tn ~ 27

References for Public Smoking 1. Gori, G.B. Hearings before the Committee on Interstate and Foreign Commerce (Subcommittee on Oversight and In- vestigations) on Cigarette Smoking and Health - Update 1978: pp. 4-58, Oct. 6, 1978. 2. Pimm, P.E., et al. Physiological Effects of Acute Passive Exposure to Cigarette Smoke. Archives of Environmental Health 33/4: 201-213, July-August 1978. 3. Hugod, C., et al. Exposure of Passive Smokers to Tobacco Smoke Constituents. International Archives of Occupa- tional and Environmental Health 42: 21-29, 1978. 4. Moser, K.M. Testimony before the Tobacco Subcommittee, Committee on Agriculture, U.S. House of Representatives, 95th Congress, Second Session, Serial No. 95-000: pp. 35-40, 258-282, Sept. 7, 1978. 5. Knoebel, S.B. Testimony before the Tobacco Subcommittee, Committee on Agriculture, U.S. House of Representatives, 95th Congress, Second Session Serial No. 95-000: pp. 49-55, 322-330,. Sept. 7, 1978. 6. Brunnemann, K.D., and Hoffmann, D. Volatile N-Nitro- samines in Tobacco Smoke and in Polluted Indoor Environ- ments. Delivered at the 2nd Joint Conference sponsored by the Chemical Institute of Canada and American Chemical Society, Montreal, Que., Canada, May 29-June 2, 1977. 7. U.S. Public Health Service. The Health Consequences Smoking. A Report of the Surgeon General: Department of Health, Education and Welfare. DHEW Pub. No. (HSM) 71-7513, 1971, 458 pp. of 1971. U.S. Washington, 8. Schrauzer, G.N. Testimony for the Subcommittee on Tobacco, Committee on Agriculture, U.S. House of Repre- sentatives, 95th Congress, Second Session, Serial No. 95-000: pp. 139-157, Sept. 7, 1978. 9. Aronow, W.S. Effect of Passive Smoking on Angina Pec- toris. New England Journal of Medicine 299/1: 21-24, July 6, 1978. 10. Aviado, D.M. Statement before Environmental Control Committee, Chicago, Ill., as entered into record of the Tobacco Subcommittee, Committee on Agriculture, U.S. Housd of Representatives, 95th Congress, Second Session, Serial No. 95-000: pp. 185-223, Sept. 7, 1978. 8

11. Booker, W.M. Testimony before the Subcommittee on Tobacco, Committee on Agriculture, U.S. House of Repre- sentatives, 95th Congress, Second Session, Serial No. 95-000: pp. 64-68, 337-345, Sept. 7, 1978. 12. Coodley, A. Letter: New England Journal of Medicine 299/16: 897, Oct. 19, 1978. 13. Fisher, E.R. Testimony before the Tobacco Subcommittee, Committee on Agriculture, U.S. House of Representatives, 95th Congress, Second Session, Serial No. 95-000: pp. 2-20, 232-371, Sept. 7, 1978. 14. Hickey, R.J. Testimony for the Tobacco Subcommittee, Committee on Agriculture, U.S. House of Representatives, 95th Congress, Second Session, Serial No. 95-000: pp. 133-135, Sept. 7, 1978. 15. Waite, C.L. Letter: New England Journal of Medicine 299/16: 897, Oct. 19, 1978. 16. Wakeham, H. Letter: New England Journal of Medicine 299/16: 896, Oct. 19, 1978. 17. Niden, A.H. No: Environmental Smoke Can Irritate, not Injure Others. Los Angeles Times, Part 5: 1, Oct. 29, 1978. 18. Rylander, R. Workshop Results: 4.1. Perspectives on Environmental Tobacco Smoke Effects. In: Environ- mental Tobacco Smoke Effects on the Non-Smoker, Report from a Workshop. University of Geneva, 1974. 19. U.S. Public Health Service. The Health Consequences of Smoking: 1975. U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (CDC) 76-8704, 1975, 235 pp. 20. Srch, M. The Significance of Carbon Monoxide During Cigarette Smoking Inside of Automobiles. Deutsche Zeitschrift fur Gerichtliche Medizin 60/3: 80-89, 1967. 21. Szadkowski, D., et al. Body Burden of Carbon Monoxide From Passive Smoking in Offices. Innere Medizin 3/6: 310-313, September 1976. 22. Cole, P.V. Comparative Effects of Atmospheric Pollution and Cigarette Smoking on Carboxyhemoglobin Levels in Man. Nature 255/5511: 699-701, 1975. 23. Anderson, G., and Dalhamn, T. Health Risks Due to Passive 29

Smoking. Lakartidningen 70/33: 2833-2836, Aug. 15, 1973. 24. Bridge, D.P., and Corn, M. Contribution to the Assess- ment of Exposure of Non-Smokers to Air Pollution from Cigarette and Cigar Smoke in Occupied Spaces. Environ- mental Research 5/2: 192-209, 1972. 25. U.S. Department of Transportation, Federal Aviation Administration. U.S. Department of Health, Education and Welfare, National Institute for Occupational Safety and Health. Health Aspects of Smoking in Transport Aircraft, Rockville, Md. AD 736097, December 1971, 85 pp. 26. Harke, H.P., et al. The Problem of Passive Smoking. II. Investigations of CO Level in the Automobile After Cigarette Smoking. Internationales Archiv fur Arbeits- medizin 33/3: 207-220, 1974. 27. Eisenbud, M., and Ehrlich, L.R. Carbon Monoxide Concen- tration Trends in Urban Atmospheres. Science 176: 193-194, April 14, 1972. 28. Sterling, T.D. Testimony before the Tobacco Subcom- mittee, Committee on Agriculture, U.S. House of Represen- tatives, 95th Congress, Second Session, Serial No. 95-000: 41-45, 282-297, Sept. 7, 1978. 29. Klosterkotter, W. State of the Medical Debate on the Theme "Passive Smoking". Delivered to conference on Passive Smoking in the Workplace, sponsored by Bavarian State Minister of Labour, Munich, Germany, March 31 - April 1, 1977. 30. Harke, H.P., and Bleichert, A. On the Problem of Pas- sive Smoking. Internationales Archiv fur Arbeitsmedizin 29: 321-322, 1972. 31. U.S. Department of Transportation, Federal Aviation Administration. Denial of Petition of the Airline Pilots Committee of 76, Public Citizen's Health Research Group and Aviation Consumer Action Project to amend Part 91, Federal Aviation Regulations: Reg. Docket No. 15614, Aug. 22, 1977. 32. U.S. Department of Transportation. News release: Peti- tion to Ban Smoking by Airline Flight Crews Denied. Sept. 1, 1977. 33. National Institutes of Health. Cigarette Smoking and 30

Airline Pilots: Effects of Smoking and Smoking Withdrawal on Flight Performance, a Report of an Expert Panel of Consultants. U.S. Department of Health, Education and Welfare. Washington, April 1978. 82-97, Sept. 7, 1978. 34. Associated Press. Nonsmoker Woes in Head or Nose? Hackensack, N.J., Sunday Record, July 10, 1977. 35. Serling, R.J. Letter: The Washington Post: July 7, 1977. A-16, 36. Terry, L.L. In: Smoke Ban Kindles Crossfire, by Frank VanDusen. The Atlantic City (N.J.) Press, Oct. wl, 1977. 37. McDougall, J.C., and Gleich, G.J. Tobacco Allergy -- Fact or Fancy? Journal of Allergy and Clinical Immun- ology 57/3: 237, 1976. 38. Salvaggio, J.E. Testimony before the Tobacco subcom- mittee, Committee on Agriculture, U.S. House of Represen- tatives, 95th Congress, Second Session, Serial No. 95-000: pp. 46-49, 298-321, Sept. 7, 1978. 39. Peshkin, M.M., and Landay, L.HJ. Cutaneous Reactions to Tobacco Antigen in Allergic and Non-Allergic Children with the Direct and Indirect (Local Passive Transfer) Methods of Testing. Journal of Allergy 10/3: 241-245, 1939. 40. Fontana, V.T., et al. Studies in Tobacco Hypersensitiv- ity III. Reactions to Skin Tests and Peripheral Vascular Responses. Journal of Allergy 30: 241-249, 1959. 41. Harkavy, J. Tobacco Allergy in Cardiovascular Disease: a Review. Annals of Allergy 26: 447-459, 1968. 42. Speer, F. Tobacco and the Non-Smoker. A Study of Sub- jective Symptoms. Archives of Environmental Health 13/6: 443-446, 1968. 43. Becker, C.G., et al. Tobacco Allergy: Immediate Cuta- neous Hypersensitivity to an Antigen Purified from Cured Tobacco and Present in Cigarette Smoke. Federation of American Societies for Experimental Biology Proceedings 44. 35/3: 673, March 1, 1976. TJ Becker, C.G., et al. gen. Proceedings of Hypersensitivity to Tobacco Anti- the National Academy of Science USA tTi O ,.-.~ 73/5: 1712-1716 May 1976. 4~, , ca ~ 31 Cn CA

45. Becker, C.G., and Dubin, T. Tobacco Allergy and Cardio- vascular Disease. Cardiovascular Medicine 3/8: 851-854, August 1978. 46. Stedman, R.L. Testimony for the Tobacco Subcommittee, Committee on Agriculture, U.S. House of Representatives, 95th Congress, Second Session, Serial No. 95-000: pp. 82-97, Sept. 7, 1978. 47. Chrambach, A., and Rodbard, D. Polyacrylamide Gel Elec- trophoresis. Science 172: 440-445, April 30, 1971. 48. Kice, J.S. Statement delivered before the American Cancer Society's National Commission on Smoking and Public Policy, Philadelphia. June 16, 1977. 49. Cameron, P., et al. The Health of Smokers' and Non- smokers' Children. Journal of Allergy 43/6: 336-341, June 1969. 50. Cameron, P., and Robertson, D. Effect of Environmental Tobacco Smoke on Family Health. Journal of Applied Psychology 57/2: 142-147, 1973. 51. Luquette, A.J., et al. Some Immediate Effects of A Smoking Environment on Children of Elementary School Age. Journal of School Health 40/10: 533-536, December 1970. 52. Colley, J.R.T. Respiratory Symptoms in Children and Parental Smoking and Phlegm Production. British Medical Journal 2: 201-204, 1974. 53. Harlap, S., and Davies, A.M. Infant Admissions to Hospital and Maternal Smoking. Lancet 1: 529-532, 1974. 54. U.S. Public Health Service. The Health Consequences of Smoking. A Report of the Surgeon General: 1972. U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (HSM) 73-8704, 1972, 158 pp. 55. Schilling, R.S.F., et al. Lung Function, Respiratory Disease, and Smoking in Families. American Journal of Epidemiology 106/4: 274-283, October 1977. 56. Schilling, R.S.F., and Bouhoys, A. Letter: British Medical Journal 3: 895, 1978. 57. Colley, J.R.T., et al. Influence of Passive Smoking and Parental Phlegm on Pneumonia and Brochitis in Early Childhood. Lancet 2: 1031-1034, 1974. 32

58. Shy, C.M., et al. The Chattanooga School Children Study: Effects of Community Exposure to Nitrogen Dioxide. Journal of the Air Pollution Control Association 20/9: 582-588, September 1970. 59. Hammer, D.I., et al. Air Pollution and Childhood Lower Respiratory Disease: Exposure to Indoor Air Pollutants Including Sidestream Smoke. 68th Air Pollution Control Association Abstracts: 157-158, June 15-20, 1975. 60. Lebowitz, M.D., and Burrows, B. Respiratory Symptoms Related to Smoking Habits of Family Adults. Chest 69/1: 48-50, January 1976. 61. Kerribijn, K.F., et al. Chronic Nonspecific Respiratory Disease in Children, Five Year Follow-Up Study. Acta Paediatrica Scandanavica Supplement 261; 1977, 71 pp. 62. Melia, R.J.W., et al. Association Between Gas Cooking and Respiratory Disease in Children. British Medical Journal 3: 149-152, 1977. 63. Langston, H.T. Statement before Environmental Control Committee, Chicago, ILL., as entered into record of the Tobacco Subcommittee, Committee on Agriculture, U.S. House of Representatives, 95th Congress, Second Session, Serial 95-000: pp. 158-184, Sept, 7, 1978. 64. Halberstam, M.J. Smoking and the Nonsmoker. New York Times Syndicate, for release May 12, 1978. 33

Smoking and Over-All Mortality "[I]t is not unreasonable to speculate that the kind of men who become regular cigarette smokers are, to a moderate degree, less inherently able to survive to a ripe old age than non- smokers." "Smoking and Health: Report of the Advisory Committee to the Surgeon General of the Public Health Service" 1964 (1) This sentiment, buried in the report released January 11, 1964, is not an unreasonable speculation today. But the mortality statistics that report developed on reported smoker/nonsmoker differences are still being used -- and abused. For instance, the HEW report to Congress of 1977-78 made the statement, based on only three population studies, The use of results from flawed population studies to frighten people by attributing large numbers of deaths yearly to smoking may be misleading and is most regrettable. that "over-all mortality rates for cigarette smokers are about 70 p ercent higher than those of nonsmokers," and discussion of smoker/nonsmoker mortality differences occupied almost half the report (2). HEW Secretary Califano took the mortality rate 35

implications, one step further, saying in January 1978 (and often through the year) that more than 300,000 persons died in 1977 from cancer and heart disease for which "smoking was major factor" (3). a For a more complete exposition of the so-called "excess deaths" concept, see Appendix. Meantime, here's how most of these sorts of figures were derived. The calcula- tions were based on the population surveys (4-12) on which the first Surgeon General's report (1) relied so heavily. Government statisticians took data from the seven surveys, conducted in various areas, in various groups of people, over varying periods, and considered for the most part only whether or not the subjects smoked. The statis- ticians then reanalyzed the data and computed what they called a mortality ratio (smokers vs. nonsmokers) of 1.68. "A mortality ratio higher than 1 implies [emphasis added] that the group of smokers has a higher over-all death rate than the non-smokers," they said in the 1964 report. Expressed another way, the mortality difference is 68 percent, and the 1977-78 report authors rounded 68 off to "about 70 percent". That's the excess percent of deaths observed over what might have been expected had smokers died at the same rate as nonsmokers. PQ ~ 0 ~ But, as the authors of the 1964 report pointed out, 4~1. w 1" Ln -o 36

the mortality difference in the seven surveys varied from 44 percent higher in British doctors (5,6) to 83 percent higher for an American Cancer Society survey conducted by volunteers (11,12). And the 1964 authors also drew attention to the possibility that bias arising from what were high nonresponse rates "might account for a mortality ratio of 1.3". This raises questions about the other biases that might affect the accuracy of the 1.68 figure. Still with us? Let's get on, then, with some of the "weaknesses" of the surveys from which the 1.7 mortality ratio was computed. Some have been enumerated by the govern- ment people themselves (1,2), some pointed out by others. But first let us consider another not unreasonable speculation by the authors of the first report: [T]he low death rates for non-smokers [lower than U.S. general death rates] suggest the possibility that the studies recruited unusually healthy groups of non-smokers (1). Design Weaknesses American Cancer Societ surveys (10-13). "Men in 25 states" was the largest of the seven surveys. The authors of the 1964 report said that this survey and the earlier "men in nine states" survey "suffer from the difficulties that the popula- tions studied are hard to define, that the smokers and non- 37

smokers were recruited by a large number of volunteer workers, and that completeness in the reporting of deaths was hard to achieve, since this depends on reports from the volunteers" (1). What they didn't say was that the larger Cancer Society survey wound up with a male lung cancer death rate twice as high as that of U.S. males nationwide and women's rate three times as high as that of U.S. females. Death rates observed for coronary heart disease and emphysema were 10 to 60 percent higher than male and female national rates (14), all of which leads to the not unreasonable speculation that the volunteers enrolled many who were already ill with these diseases. If persons with alleged smoking-related ills were overrepresented in the ACS population, so were residents of coastal and urban areas, including the industrialized North- east; the mountain states and the Northwest tier were excluded entirely. There was an overabundance of better-educated, native-born, Protestant whites and a dearth of blacks (14). And a map of the ACS survey states includes almost all the so- called lung cancer hot spots identified in the National Cancer Institute's new "cancer atlas" (15) (see chapter on Cancer in the Work Place). tJ cn 0 ~ P One finding from the later ACS survey is surprising w . , a* and disturbing if one believes that tobacco causes disease: 38

Men who smoked cigars only -- as well as those who smoked pipes only -- had lower mortality ratios than those who smoked no tobacco at all (1). This finding can not easily be dismissed as a quirk of nature because the British doctors study (5,6) also reported this "incongruity". The U.S. veterans survey (7,16). Just as nonrepresentative of the total U.S. population as the ACS's predominantly middle and upper socioeconomic subjects, the veterans were mostly white-collar, skilled workers who served in World War I. Their smoking habits, unfortunately, were recorded only when they first returned their questionnaires (in two waves,in 1954 and 1957). Reporting on later data from the same population (17), the HEW 1977-78 report to Congress noted as a weakness "the lack of information about more recent changes in smoking habits" (2). One of the strange findings: Veterans who had smoked cigarettes and other products for 25 to 35 years had lower mortality ratios than those smoking only 15 to 24 years (1). British doctors survey (5,6,18). An obvious drawback in this survey is that its subjects were highly selected, British professional men who shouldn't be considered comparable to any general U.S. population. More importantly, the survey actually showed that quitting smoking did not reduce mortality -- the exact opposite finding from that claimed by many who have cited e s 39

it as proof that cigarettes shorten lives. When a risk factor has been associated in a popula- tion with increased incidence of a disease (or death), removal of that risk factor should result in a drop in that disease (or death) in that population. A 50 percent reduction in cigarette smoking between 1951 and 1965 did'not change the over-all death rates in those years in the British doctors (19)1 And there went the remove-the-risk-factor-and-reduce-the-risk theory-- at least as far as the British doctors were concerned. Another problem with the physicians survey was referred to briefly before, that of "nonresponse bias". Only two-thirds of the doctors who were sent questionnaires replied (1). In other words, one in every three British doctors con- tacted either was not interested enough to return his ques- tionnaire (or had some reason not to) or else gave answers that were rejected as incomplete. There is no way of knowing if those who do not participate in a survey of this sort are similar to those who do. The authors of the British doctors survey wrote later that a sample of their nonrespondents indicated they differed in several respects from a sample of respondents, and the authors felt "sure that the doctors who chose to answer were not rep- resentative of the total" (18). That such nonresponse can N tn 0 seriously bias this sort of investigation and lead to a "spate 4N. P of doubts" h as been expressed eloquently by one of the euthors ~ o• w 40

of the same survey of British doctors (20). It is not surprising that the authors of the 1964 report properly warned that "none of the populations was designed, in particular, to be representative of the U.S. male population". Or that they continued: "Any answer to the ques- tion 'to what general population of men can the results be, applied?' must involve an element of unverifiable judgment... The seven studies differ considerably in size. They vary also in the extent to which they are free from methodologic weak- ness" (1). Almost a decade later a Canadian researcher commented on the methodological problems in the surveys before an American Statistical Association meeting. He included a com- plaint that "few if any of the many studies reporting a link between smoking and disease have ever been published in a principal statistical journal where the methods of sampling and data analysis would have received adequate review" (14). In . view of the many critical defects in these surveys, he sug- gested that researchers should reevaluate their past reliance on them. Association vs. Cause Long forgotten in any claims of high mortality ratios andlor excess deaths in smokers is the careful caveat written 41

into the first Surgeon General's report by its authors. They said: Statistical methods cannot establish proof of a causal relationship in an association. The causal significance of an association is a matter of judg- ment that goes beyond any statement of statistical probability (1). Many who have opposed smoking appear to infer causality from association in all causes of death (including accidents and suicides) which have reported smoker/nonsmoker mortality ratios larger than that magic number 1. That nation- al abstention from smoking would automatically prolong the lives of those who had smoked is a most unreasonable specula- tion. And that's exactly what a well-respected government health statistician told an American Cancer Society meeting recently: [M]ost of the large-scale studies on smoking and health have tended to investigate the role of smoking independent of other behavioral variables, such as alcohol consumption and other lifestyle factors, occupational and environmental hazards and certain psychological factors. These variables are known to be related to health status... Thus it may well be that the elimination of smoking without any changes in the other factors will have only a partial impact on health status (emphasis added)(21). =42

Conclusion: Then and Now The past reliance on population surveys to indict tobacco in disease causation is open to severe criticism because of the many inherent weaknesses in these studies. Even outspoken opponents of smoking have recognized this: "As mentioned previously, the smokers and non-smokers in th ese studies may differ with respect to other variables that might influence the death rate." "Smoking and Health: Report of the Advisory Committee to the Surgeon General of the Public Health Service" 1964 (1) "Blanket assumptions that every disease associated with smoking is caused by smoking create a credibility gap..." Thomas G. Vogt, M.D., M.P.H.(22) 43

References for Smoking and Over-All Mortality 1. U.S. Public Health Service. Smoking and Health. Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington, U.S. Department of Health, Education and Welfare, PHS Pub. No. 1103, 1964, 387 pp. 2. U.S. Public Health Service. The Health Consequences of Smoking: 1977-78. U.S. Department of Health, Education and Welfare. Washington, July 27, 1978, 73 pp. 3. Califano, J.A. Address of the Secretary of Health, Education and Welfare before the National Interagency Council on Smoking and Health, Washington, D.C., Jan. 11, 1978. 4. Best, E.W.R., et al. A Canadian Study of Mortality in Relation to Smoking Habits, a Preliminary Report. Canadian Journal of Public Health 52: 99-106, 1961. 5. Doll, R. Personal communication to the Surgeon General's Advisory Committee on Smoking and Health (1). 6. Doll, R., and Hill, A.B. Lung Cancer and Other Causes of Death in Relation to Smoking. British Medical Journal 2: 1071-1081, 1956. 7. Dorn, H.F. The Mortality of Smokers and Non-Smokers. Proceedings of the Social Statistics Section, American Statistical Association 34-71, 1958. 8. Dunn, J.E., et al. Lung Cancer Mortality Experience of Men in Certain Occupations in California. American Journal of Public Health 50: 1475-1487, 1960. 9. Dunn, J.E., Jr., et al. California State Department of Public Health. Special Report to the Surgeon General's Advisory Committee on Smoking and Health (1). 10. Hammond, E.C., and Horn, D. Smoking and Death Rates -- Report on 44 Months of Follow-up on 187,783 men. Part I. Total Mortality. Part II. Death Rates by Cause. Journal of the American Medical Association 166: 1159-72, 1294-1308, 1958. 11. Hammond, E.C. Special Report to the Surgeon General's Advisory Committee on Smoking and Health (1). 12. Hammond, E.C. Special Report to the Surgeon General's Advisory Committee on Smoking and Health (1). 44

13. Hammond, E.C. Smoking in Relation to the Death Rates of One Million Men and Women. In: Haenszel, W., editor. Epidemiological Approaches to the Study of Cancer and Other Diseases. Bethesda, Md., U.S. Public Health Service, National Cancer Institute Monograph 19, January 1966, pp. 127-204. 14. Sterling, T.D. The Effect of Self-Selection Factors in the Study of Smoking and Lung Cancer. Presented to the Bio- metrics Section on Statistical Aspects of Population Research of the American Statistical Association, Aug. 16, 1972. 15. Mason, T.J., et al. Atlas of Cancer Mortality for U.S. Counties: 1950-1969. National Cancer Institute, U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (NIH) 75-780, 1976, 103 pp. 16. Kahn, H.A. The Dorn Study of Smoking and Mortality Among U.S. Veterans: Report on Eight and One-Half Years of Observation. In: Haenszel W., editor. Epidemiological Approaches to the Study of Cancer and Other Diseases. Bethesda, Md., U.S. Public Health Service, National Cancer Institute Monograph 19, January 1966, pp. 1-27. 17. Rogot, E. Smoking and General Mortality Among U.S. Veterans 1954-1969. U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (NIH) 74-544, 1974, 65 pp. 18. Doll, R., and Hill, A.B. Mortality in Relation to Smoking: Ten Years' Observations of British Doctors. British Medical Journal 1: 1399-1410, 1964. 19. Seltzer, C.C. Critical Appraisal of the Royal College of Physicians' Report on Smoking and Health. Lancet 1: 243-248, 1972. 20. Hill, A.B. Statistical Methods in Clinical and Preventive Medicine. E. and S. Livingstone Ltd., Edinburgh and London, 1962. 21. Wilson, R.W. Statement befor,e the American Cancer Soci- ety's National Commission on Smoking and Public Policy, Philadelphia, June 16, 1977. 22. Vogt, T.M. In: Jarvik, M.E., et al, editors. Research on Smoking Behavior. Rockville, Md., U.S. Public Health Service, National Institute on Drug Abuse Monograph 17, December 1977, pp. 228-229. 45

Women and Smoking After two decades of denouncing smoking and claiming "proof" that cigarettes cause various diseases and disorders in men, anti-smoking organizations have in recent years launched special campaigns to persuade women that they, too, are ad- versely affected by cigarettes. Their alarums usually begin with the charge that the woman who smokes in pregnancy may harm her infant. There are claims, too, of reported increases in lung cancer mortality in women as a result of their smoking. Because of this new emphasis on the ladies, we devote a Inconsistent findings from studies of smoking women and their children make it impossible to draw con- vincing conclusions from the data. chapter here to what HEW has called "the smoking-related problems unique to women" (1) -- and a look at some of the unexplainable lung cancer mortality trends for women. Pregnancy Outcome A sizable section of the HEW 1977-78 report to Con- gress on smoking and health (1) was devoted to smoking women N U1 and their pregnancies. It concluded with the strong language a that cigarette smoking was "probably causally associated" W N %o t 47

with higher late fetal and infant mortality. However, the actual data mainly relied upon by the HEW authors in reaching this conclusion suggest that any relationship between maternal smoking and pregnancy outcome is far from clear and any claims of causality have highly questionable foundation. The study relied upon by HEW in that last report to Congress was a retrospective analysis at Johns Hopkins of 51,490 births recorded in 10 teaching hospitals in 1960 and 1961 (2). Data were collected on infant birth weight, infant mortality, prematurity and placental complications. The statistical analyses of the data on infant mortality indicated that a history of a previous pregnancy loss, the mother's hospital status (private or public patient -- a socioeconomic indicator) and a variable related to age and number of previous pregnancies had "greater effects" on perinatal mortality than maternal smoking level. In their analyses for prematurity and placental complications, the researchers found that previous pregnancy loss and hospital pay status were more strongly related to unfavorable outcomes than maternal smoking level. The reported importance of previous pregnancy history and hospital pay status strongly indicates that a mother's N pregnancy experience may well be determined by who the mother ~ ~ is -- her constitution or innate characteristics -- rather 4 Lit ~ than whether or not she smokes. The `complexity of all of ~ 48

these findings and the areas to which the data point for further research seem strikingly inconsistent with the unswerv- ing and exclusive emphasis on the mother's smoking habits. Low-Birth-Weight Babies Like most other pregnancy studies, the work at Johns Hopkins found that smoking women on average have smaller infants than nonsmokers -- more of what are called low-birth- weight (LBW) babies. LBW infants weigh 2,500 grams -- about 5.5 pounds -- or less. Why and how this happens has not been explained. But the possibility that a common factor predisposes women both to smoke and to have a higher proportion of LBW infants was recently described by the director of a child health study who suggested that "the smoker and not the smoking" may determine whether a woman has an LBW infant (3). Dr. Bea van den Berg took over direction of the large PHS- funded California study from the late Dr. Jacob Yerushalmy, who f irst proposed, as early as 1964, the hypothesis that a mother's smoking may serve as a marker for -- but not as a causal factor in -- the birth of LBW infants (4). Yerushalmy contended that ineffective randomization a nd the problem of self-selection in studies comparing smoking and nonsmoking mothers made it difficult to draw any infer- ences from the observation that smokers seem to have more LBW 'infants (5). In perhaps his best-known study, he identi- 49

fied a group of women who began to smoke after their first children were born (6). Comparing the birth weights of chil- dren born before and after the women began smoking, he dis- covered that both groups of children were lighter than the children of nonsmoking mothers. He said this indicated some women will have smaller infants whether or not they smoke. Two reseachers published data in 1977 that appear to support Yerushalmy's hypothesis. A National Institutes of Health epidemiologist found that differences in mean birth weights of infants born to women who smoked during one pre- gnancy but not another were "more consistent with the self- selection hypothesis" than the causal hypothesis (7). An Australian who worked with records of 1,200 maternity patients concluded his findings were "compatible" with the theory that maternal smoking does not cause LBW but is "an index" of some other factor or factors (8). Perinatal MortalitV, Any claim that maternal smoking during pregnancy is causally related to increased perinatal mortality is not supported by the scientific evidence. Yerushalmy, for example, found that the mortality rate of LBW infants was considerably lower for those with smoking mothers than for those with nonsmoking mothers (9). He contended that his data argued against the proposition that cigarette smoking acts as an external factor that interferes with fetal development. 50

In 1978, the editor of the British journal, Public Health, wrote that evidence that small infants of smoking mothers do not share the high mortality of infants of the same weight born to nonsmoking mothers "has been disregarded." He suggested, "We may tell women that if they smoke their baby may be small. But [we] should not claim risk to life" (10) . Spontaneous Abortion In the 1973 HEW report to Congress, the last specifi- cally to discuss spontaneous abortion, the authors said several studies had reported finding a significantly higher, dose- related incidence among cigarette smokers. But they conceded that "the lack of control of significant variables other than cigarette smoking does not permit a firm conclusion to be drawn about the nature of the relationship" (11). No firm conclusion about the "nature" of the rela- tionship can be drawn now, either. A recent study by New York researchers did assert that smoking is "a risk factor" for spontaneous abortion (12). However, the researchers found no statistically significant relationship between the amount smoked and the rates of spon- taneous abortion. Moreover, their emphasis on certain data in the study was criticized by another researcher, who said this focus magnified "the apparent effect of smoking" in the higher- 51

risk age groups -- the younger and older mothers-to-be (13). He suggested that if some of the women were smoking "because they were uptight about a floundering pregnancy" that fact "might distort the picture just enough to make it appear that smoking is an etiologic agent of spontaneous abortion, when in fact it may merely be a more prevalent behavior characteristic in a troubled pregnancy." Failure to consider this and a number of other factors caused him to conclude, he said, that "we are still at a loss for the cause of spontaneous abortion." That smoking is a risk factor for spontaneous abortion is not supported by other studies, which have failed to show any significant link with smoking. These include two published since 1976 (14, 15). Another, conducted in Sweden, examined a variable that is not always considered. It found that an overall increased risk of spontaneous abortion among smoking women was almost completely due the pregnancy was unwanted (16). to the fact that A British Medical Journal editorial of less than a year ago puts the reported relationship in perspective more succinctly than anything we could say: What remains to be established is whether the association between cigarette smoking and spon- taneous abortion is causal...Only by identifying a mechanism by which cigarette smoking could give rise to spontaneous abortion could we be confident of a causal relation (17). 52

Congenital Malformation A physician appearing before an American Cancer Society "forum" on smoking stirred the audience with his' charge that smoking is "likely to cause birth defects" (18). However, his opinion was not shared by another physician, appearing at a similar ACS "forum" two weeks later. The second doctor said: ...I don't think anyone has identified absolute evidence that this [congenital malformation] is a result of the chronic or even acute smoking of the mother (19). These conflicting opinions, especially within one anti-smoking organization, reflect the inconclusive scientific findings in this area. Several large-scale population studies have failed to establish a relationship between smoking and congenital malfor- mation (9, 15, 20, 21). Another, examining congenital mal- formation diagnosed during the first five years of life, found that fewer such conditions occurred in children born to women who smoked during pregnancy than to women who never smoked or to women who stopped at some time before becoming pregnant (22). E ven the New York researchers who reported an association between smoking and spontaneous abortion concluded, after study of the scientific literature on smoking and infant malforma- ~ ` 53

tion, "it is unlikely that smoking acts to cause fetal anoma- lies" (12). That emotionalism c.an override objective analysis in any area of pregnancy and childbirth is illustrated by the headlines which accompanied the release of a study by a Pitts- burgh pathologist who claimed maternal smoking was related to congenital malformation (23). Although one headline read "Baby Brain Defect Linked to Smoking" (24), examination of the research paper revealed that the pathologist had described this finding only as an "apparent association" that "requires further analysis." This sort of proviso, of course, never appears in headlines. Child Development Another favorite claim of anti-smokers is that smoking during pregnancy retards the subsequent growth and learning ability of the child. In fact, HEW Secretary Califano in early 1978 spoke of the "developmentally disabled" children of smoking mothers (25). The basis for these allegations? Apparently it is data from an on-going perinatal and child follow-up study in Britain which indicated that the children of smoking mothers lagged behind the children of nonsmoking mothers in physical a nd mental development (26-28). The authors did note that the 54

effect of smoking during pregnancy is "relatively small" in comparison with the effects of some other factors, such as social class and the number of older and younger children in the household (28). In the British study, the children of smoking mothers were on the average 1 centimeter -- or only about three-tenths of an inch -- shorter than children of nonsmoking mothers (26). There was also a four-month difference in reading ability between the two groups of children (27). But analysis of physical growth showed that a number of other factors were associated with size at age 7. For example, the child of a blue-collar family was on the average 1.3 centimeters shorter than the child of wealthier parents, while the fourth-born child was usually 2.3 centimeters shorter than the first-born. In a later report from the same British study, researchers examined the children at age 11 and measured only minor differences in either height or mental development of children born to smoking and nonsmoking mothers (28). They also reported that these differences were less than the effects of some of the other factors considered. For instance, the difference between a child from a household with no older children and one from a household with three or more was, on the average, 16 months for general ability, 29 months for reading, 14 months for mathematics, and 4 centimeters for height. 55

The results of the British study were inconsistent with those of Johns Hopkins researchers who followed children born of smoking and nonsmoking mothers in 1962 and 1963. They reported: "At four and seven years there was no significant difference in either physical measurements or intellectual functioning" (29). Despite such studies, there are still those who look only at whether the mother smoked during pregnancy to explain possible differences in children's growth and learning skills. This shortcoming was apparent in the 1977-78 HEW report to Congress on smoking and health (1). The authors discussed a California study in which children of smoking mothers were found to be shorter at age five than children of nonsmoking mothers (30). But the HEW writers didn't indicate that the study showed a difference of only 0.9 centimeters, and that the researchers had attributed nearly 90 percent of the varia- tions to "parental stature alone. n Only 2 percent of the .9 centimeter difference could possibly be due to association with smoking -- about 7 one-thousandths of an inch -- according to the California investigators. N Oral Contraceptives p ~ ~ ~ W i" The scientific literature does not support the claim ~ that oral contraceptive users who stop smoking significantly decrease their disease rates. In other words, there is no 56

scientific basis for any assurance that oral contraceptive users are not likely to develop certain diseases if they stop smoking. The new warning insert now required by the Federal' Food and Drug Administration in oral contraceptive packages in effect implies such assurance. And the FDA came under attack at a Congressional hearing last fall for its decision requiring the new inserts. Several prominent statisticians who had examined the studies relied upon by the agency in its decision testif ied (31-34). One noted that: The warning as stated implies that there is a causal effect of cigarette smoking on the incidence of cardiovascular disease. No statistical study can establish causality. At best it can establish a high probability of possible interrelationships (33). The studies which the statisticians criticized claimed to have shown that smokers who use oral contraceptives have an increased risk of cardiovascular disease (35-38). Their censure, the statisticians said, was based on statistical weaknesses, inadequate sample size and other methodological problems in the studies. Some of these weaknesses had been noted by the authors of the original reports: These estimates of risk...still need to be inter- preted with caution, as a number of assumptions have necessarily had to be made in their calculation and the margin of error is likely to be fairly wide (35). It is essential to point out that the mortality estimates used in this paper are based on small r 57

numbers and may be subject to large sampling errors. These estimates are also subject to upward and downward biases, which may not cancel each other out... The net effect of these various factors cannot be estimated without additional research, preferably in different settings (37). These estimates [death-rates) are based on small numbers and are necessarily approximate. Without more data it is not possible to examine the inter- relationships of age, smoking, and duration of oral contraceptive use...(38). Several witnesses expressed concern that a statisti- cal relationship between oral contraceptive use, smoking and illness may influence scientists to overlook other factors that may explain the observed relationship. Two other researchers urged in the British journal Lancet that scientists not let "preconceived ideas affect objectivity" in examination of other possible hypotheses. Left unanswered in research so far, they wrote, is the con- founding question of self-selection (39). And they asked: Is it possible that the women who smoke and use oral contraceptives (particularly the older group) are simply reflective of a more flamboyant lifestyle which may well include more stress, more medication (including "downers" or marijuana use?...There seems to the data which would answer question -- is it the smoker more alcohol, and "uppers"), be nothing in once and for all the or the smoking which creates risk difference and is it the ceptive user or oral-contraceptive at the bottom of it all? oral contra- use which is They concluded that "somehow epidemiological and laboratory studies must be designed to distinguish among the 58

possible aetiologies." What they called "the mix of factors", which include genetics, they wrote, "may be more complex than we think." Smoking and Early Menopause In a widely publicized study which appeared in 1977, Boston researchers reported they had found that smokers undergo menopause earlier than nonsmokers (40). According to some voluntary health associations and anti-smokers, this appeared to present a health hazard, because onset of change of life supposedly deprives a woman of the "protection" she is b elieved to have from female sex hormones (estrogens) against cardiovascular disease (CVD) in her reproductive years. However, this claim can be disputed for several reasons. First, the CVD vital statistics don't support such a theory. If women actually lost some form of hormonal protec- tion with the onset of menopause, reported CVD death rates presumably would jump during the menopausal age span -- rather than continue the expected steady increase. An editorial in Lancet examined this question a few months after publica- tion of the Boston research. And it pointed out: Mortality statistics do not seem to support the suggestion that the menopause has any effect on the risk of CHD [coronary heart disease], since the death-rate from the disease increases steadily with advancing age (41). 59

Two researchers who studied trends in CHD mortality data for women in England and Wales concluded that women do not lose protection from CHD after the menopause (42). hardening of the arteries (43). The researchers foun•d that the women did not have estrogen deficiencies, as might be expected in the hormonal protection theory. Instead, they found that a family history of coronary disease, hypertension or diabetes was "the most consistent single factor" found among the women. Finally, a study of women who underwent surgery for removal of their ovaries -- a procedure resulting in estrogen deprivation -- reported that they had no more coronary artery disease than an age-matched control group (44). The authors wrote that "our data support previous reports which question the protective effect of estrogen on development of atherosclerotic coronary artery disease." Among the facts that are not mentioned by those claiming a smoking-early menopause link is that other variables h ave been associated with the onset of menopause. For in- s tance, a Public Health Service s tudy of 1,200 women f ound that the median menopausal age tended to be lower in black women and in white women from lower income brackets and rural areas (45). 60

The PHS study also pointed to a fact that has been discussed in other studies: that lean women undergo menopause slightly earlier, and the thinner their measured skinfolds, the earlier the change in their menses. The leanness factor also was reported by other researchers in a 1976 study (46). And what everyone has forgotten is that women who smoke have been found to be leaner than women who do not (47-49). So one is left to wonder, if leaner women experience earlier menopause, is it because they are leaner, or because they smoke, or because of the kind of persons they are? Women and Lung Cancer Some persons who disapprove of cigarette smoking claim that the larger number of women smoking accounts for their rising lung cancer death rates. This claim does not stand up well under critical scrutiny. A major weakness in the claim can be seen in a comparison of international lung cancer patterns. Lung cancer death rates for U.S. women have reportedly been rising faster year to year than those in men since the early 1960s (50). The U. S. situatio,n is drastically different from 61

that in Europe. In a World Health Organization report, a professor of actuarial science reported in 1977 that while men's lung cancer death rates had risen relatively steeply in five European countries, none of the women's rates had risen significantly. In fact, the over-all lung cancer mor- tality in French women had dropped (51). In the United Kingdom declining rates of increase have been observed in younger women; simultaneously, a downhill trend in male rates has been reported in the older age groups (52). Thus, lung cancer occurrence patterns in women differ in various countries and they are also quite dissimilar from those of males. Even when allowance is made for the later popularity of smoking among women, there is still no consistent trend of increasing lung cancer mortality rates. These data are not at all compatible with the contention that the cause of lung cancer in both sexes is the same agent -- cigarette smoking. Some scientists believe that the recent rise in lung cancer in women is more artificial than real, because physicians order diagnostic tests more frequently now for women patients they know to be smokers. For example, a doctor at the Yale Medical School reported that over a 12-year period hospital records indicated a dramatic increase in the use of sputum smear tests for women lung cancer patients. He commented: "This increase in the search rate for women may 62

possibly play a role in various recent reports of rising rates of lung cancer in women" (53). Therefore, it is highly likely that more lung cancer among women is being diagnosed because of the more frequent application of diagnostic techniques than in the earlier years -- when many cases might not have been so diagnosed. Often overlooked in the hurried attempts to put the full blame for women's lung cancer on smoking is the fact that the proportion of cases of this disease with the cell type that has been most frequently associated with smoking has changed very little in the past 25 years (54,55). This raises the question of why researchers haven't observed an immense increase in squamous-cell lung cancers if the causal agent is cigarette smoking. Simple explanations for this so-called "epidemic" of lung cancer in women based solely on smoking habits are clearly inadequate. Recent studies have suggested that industrial or occupational exposures may be important in lung cancer causa- tion in women. A study in Los Angeles County reported an increased risk of lung cancer in women who worked as beauti- cians, assemblers and waitresses (56). Increased lung cancer mortality rates for women were reported in counties in the U.S. in which certain heavy industries are located. In some, the rates for women were more than the general U.S. rate (57). a third higher than k i4l 63

The need for further occupational, geographic and socioeconomic studies of women and disease should be recognized from these and similar research findings. A single-minded focus on smoking hampers the search for lung cancer causation. 64

References for Women and Smoking 1. U.S. Public Health Service. The Health Consequences of Smoking: 1977-78. U.S. Department of Health, Educa- tion and Welfare. Washington, July 27, 1978, 73 pp. 2. Meyer, M.B., et al. Perinatal Events Associated with Maternal Smoking During Pregnancy. American Journal of Epidemiology 103/5: 464-476, May 1976. 3. van den Berg, B.J. In: LBW In Infants of Smokers Not Related Just to Smoking Itself, by Annette Oestreicher. Ob. Gyn. News 13/4: 1, 27, Feb. 15, 1978. 4. Yerushalmy, J. Mother's Cigarette Smoking and Survival of Infant. American Journal of Obstetrics and Gynecology 88/4: 505-518, Feb. 15, 1964. r ?f _ 5. Yerushalmy, J. Self-Selection -- A Major Problem in Observational Studies. Proceedings of the 6th Berkeley Symposium on Mathematical Statistics and Probability, Vol. 4: 329-342, 1972. 6. Yerushalmy, J. Infants with Low Birth Weight Born Before Their Mothers Started to Smoke Cigarettes. American Journal of Obstetrics and Gynecology 112/2: 277-284, Jan. 15, 1972. 7 Silverman, D.T. Maternal Smoking and Birth Weight. American Journal of Epidemiology 105/6: 513-521, June 1977. 8 Donovan, J.W. Randomised Controlled Trial of Anti- Smoking Advice in Pregnancy. British Journal of Preven- tive and Social Medicine 31: 6-12, 1977. 9 Yerushalmy, J. The Relationship of Parents' Cigarette Smoking to Outcome of Pregnancy -- Implications as to the Problem of Inferring Causation from Observed Associa- tions. American Journal of Epidemiology 93/6: 443-456, June 1971. 10. Robertson, J.S. Health and Education -- Smoking. Public Health 92/2: 54-55, 1978. 11. U.S. Public Health Service. The Health Consequences of Smoking: 1973. U.S. Department of Health, Education N tn O P and Welfare. Washington, DHEW Pub. No. (HSM) 73-8704, - 4_1Z 1973, 249 pp. w ~ m 65

12. Kline, J., et al. Smoking: A Risk Factor for Spontaneous Abortion. New England Journal of Medicine 297/15: 793- 796, Oct. 13, 1977. 13. McKean, H.E. Letter: New England Journal of Medicine 289/2: 113, Jan. 12, 1978. 14. Alberman, E., et al. Maternal Factors Associated with Fetal Chromosomal Anomalies in Spontaneous Abortions. British Journal of Obstetrics and Gynecology 83: 621-627, August 1976. 15. Rummler, V. S. The Effect of Cigarette Smoking on the Weight and Length of the New Born. Z. Arztl. Fortbild. 71: 293-297, 1977. 16. Kullander, S . , and Kallen, B. A Prospective Study of Smoking and Pregnancy. Acta Obstetrica et Gynecologica Scandanavica 50/1: 83-94, 1971. 17. Editorial. Cigarette Smoking and Spontaneous Abortion. British Medical Journal 1: 259-260, 1978. 18. Ryser, H.J.-P. Statement before the American Cancer Society's National Commission on Smoking and Public Policy, Boston, June 2, 1977. 19. Allen, J.E. Statement before the American Cancer Soci- ety's National Commission on Smoking and Public Policy, Philadelphia, June 16, 1977. 20. Ontario Department of Health. Second Report of the Perinatal Mortality Study in Ten University Teaching Hospitals. Toronto, Canada, Ontario Department of Health, Ontario Perinatal Study Committee, Vol. 1, 1967, 275 pp. 21. Hollingsworth, D.R., et al. Abnormal Adolescent Primi- parous Pregnancy: Association of Race, Human Chorionic Somatomammotropin Production, and Smoking. American Journal of Obstetrics and Gynecology 126/2: 230-237, 1976. 22. van den Berg, B.J. Epidemiologic Observations of Pre- maturity: Effects of Tobacco, Coffee and Alcohol. In: Reed, D.M., and Stanley, F.J., editors. The Epidemiology of Prematurity. Urban & Schwarzenberg, Baltimore -- Munich, 1977, pp. 157-176. 23. Naeye, R.L. Relationship of Cigarette Smoking to Congen- ital Anomalies and Perinatal Death: A Prospective 66

Study. American Journal of Pathology 90/2: 289-294, February 1978. 24. Naeye, R.L. In: Baby Brain Defect Linked to Smoking, by Dolores Frederick. The Pittsburgh Press, March 12, 1978. 25. Califano, Education J.A. Address by the Secretary of Health, and Welfare before the National Interagency Council on Smoking and Health, Washington, D.C., Jan. 11, 1978. 26. Goldstein, H. Factors Influencing the Height of Seven Year Old Children -- Results from the National Child Development Study. Human Biology 43: 93-111, 1972. 27. Davie, R., et al. From Birth to Seven: The Second Report of the National Child Development Study (1958 Cohort). Longman in association with the National Children's Bureau, London, 1972. 28. Butler, N.R., and Goldstein, H. Smoking in Pregnancy and Subsequent Child Developmen t. British Medical Journal 4: 573-575, 1973. 29. Hardy, J.B., and Mellits, E.D. Does Maternal Smoking During Pregnancy Have a Long-Term Effect on the Child? Lancet 2: 1332-1336, 1972. 30. Wingerd, J . , and Schoen, E.J. Factors Influencing Length at Birth and Height at Five Years. Pediatrics 53/5: 737-741, May 1974. 31. Kastenbaum, M.A. Hearing before a Subcommittee of the Committee on Government Operations, U.S. House of Repre- sentatives, 95th Congress, Second Session. Quality of Scientific Evidence in FDA Regulatory Decisions (The Adoption of an Antismoking Warning in Oral Contraceptive Pill Labeling). Oct. 4, 1978, pp. 2-14, 30-35. 32. Cox, G.M. Hearing before a Subcommittee of the Committee on Government Operation s, U.S. House of Representatives, 95th Congress, Second Session. Quality of Scientific Evidence in FDA Regulat ory Decisions (The Adoption of an Antismoking Warning in Oral Contraceptive Pill Labeling). Oct. 4, 1978, pp. 15-29. 33. Gibbons, J.D. Hearing bef ore a Subcommittee of the Committee on Government O per ations, U.S. House of Repre- sentatives, 95th Congress, Second Session. Quality of R1 tn 0 Scientific Evidence in FDA Regulatory Decisions (The ~ Adoption of an Antismoking Warning in Oral Contraceptive ~ 67

Pill Labeling). Oct. 4, 1978, pp. 35-132. 34. Budne, T.A. Hearing before a Subcommittee of the Commit- tee on Government Operations, U.S. House of Representa- tives, 95th Congress, Second Session. Quality of Scienti- fic Evidence in FDA Regulatory Decisions (The Adoption of Antismoking Warning in Oral Contraceptive Pill Label- ing). Oct. 4, 1978, pp. 132-138. 35. Mann, J.I., and Inman , W.H. Oral Contraceptives and Death from Myocardial Infarction. British Medical Journal 2: 245-248, 197 5. 36. Mann, J.I., et al. O ral Contraceptive Use in Older Women and Fatal Myocardial Infarction. British Medical Journal 2: 445-447, 1976. 37. Jain, A.K. Mortality Risk Associated with the Use of Oral Contraceptives. Studies in Family Planning 8/3: 50-54, March 1977. 38. Beral, V. Mortality Among Oral-Contraceptive Users. Lancet 2: 727-731, 1 977. 39. Haack, D.G., and Mc Kean, H.E. Letter: Lancet 2: 1024, 1977. 40. Jick, H.J. , et al. Relation Between Smoking and Age of Natural Menopause. Lancet 1: 1354-1355, 1977. 41. Editorial. Coronary Heart Disease in Young Women. Lancet 2: 282-283, 1977. 42. Heller, R.F., and Jacobs, H.S. Coronary Heart Disease in Relation to Age, Sex, and the Menopause. British Medical Journal 1: 472-474, 1978. 43. Engel, H.J., et al. Coronary Artery Disease in Young Women. Journal of the American Medical Association 230/11: 1531-1534, Dec. 16, 1974. 44. Manchester, J.H., et al. Premenopausal Castration and Documented Coronary Atherosclerosis. American Journal of Cardiology 28: 33-38, July 1971. 45. MacMahon, B., and Worchester, J. Age at Menopause: U.S. 1960-1962. Public Health Service, U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (HSM) 73-1268, October 1966. 1V cn 0 ~ 46. Hjortland, M.C., et al. Some Atherogenic Concomitants ~ 41. of Menopause: The Framingham Study. American Journal W 68

of Epidemiology 103/3: 304-311, September 1976. 47. Bjelke, E. Variation in Height and Weight in the Nor- wegian Population. British Journal of Preventive and Social Medicine 25: 192-202, 1971. 48. Higgins, M.W. and Kjelsberg, M. Characteristics of Smokers and Nonsmokers in Tecumseh, Michigan. II. The Distribution of Selected Physicial Measurements and Physiologic Variables and the Prevalence of Certain Diseases in Smokers and Nonsmokers. American Journal of Epidemiology 86/1: 60-77, July 1967. 49. Kopczynski, J. Height and Weight of Adults in Cracow. III. Weight by Age and Cigarette Smoking. Epidemiology of Noninfectious Diseases. Epidemiological Review 26/4: 452-464, 1972. 50. Schneiderman, M.A. , and Levin, D.L. Epidemiology of Cancer and Tobacco Use: Trends and Trend Indicators. In: Proceedings of the 3rd World Conference on Smoking and Health, 1975, Vol. II. Public Health Service, U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (NIH) 77-1413, 1977, pp. 73-84. 51. Benjamin, B. Trends and Differentials in Lung Cancer Mortality. World Health Statisti cal Report 30/2: 118-128, 1977. 52. Adelstein, A.M. Current Vit al Statistics: Methods and Interpretation. British Me dical Journal 2: 983-987, 1978. 53. Feinstein, A.R., and Wel ls, C.K. Cigarette Smoking and Lung Cancer: in Epidemiologic The Rates Pr of oblems of "Detection Bias" Disease. Clinical Research 22/3: 555A, April 1974. 54. Harley, H.R.S. Cancer of the Lung in Women. Thorax 31/3: 254-264, June 1976. 55. Kennedy, A. Relationship Between Cigarette Smoking and Histological Type of Lung 28/2: 204-208, March 1973. Cancer in Women. Thorax 56. Menck, H.R., et al. Lung Cancer Risk Among Beauticians and Other Female Workers: Brief Communication. Journal of the National Cancer Institute 59/5: 1423-1425, November 1977. N CA 0 ~ 57. Blot, W.J., and Fraumeni, J.F., Jr. Arsenical Air -p Pollution and Lung Cancer. Lancet 2: 142-143, 1975. ~ w N 69

Cancer in the Work Place After HEW Secretary Califano announced before an AFL-CIO audience in September 1978 that at least 20 percent. of all cancer in the U.S. may be work-related (1), a news weekly for health professionals began its report on the speech like this: HEW has leaped into the midst of the continuing dis- agreement over just how much cancer is caused by workers' exposure to carcinogens on the job (2). Previous estimates had ranged only from 1 to 5 percent, and there were immediate complaints after the Secre- tary's talk from business interests, the American Cancer Society and some scientists (2-12). The reason for the ACS complaint was not immediately discernible. Its national president said only that the f igure was "totally out of line, much too high" (9) . The The almost exclusive focus on individual smoking habits in the study of disease may have delayed needed research into possible occupational and en- vironmental causes. N concern of business and industrial leaders, fighting ever- ~ 0 ~ increas`ing government regulation in the work place, was N %o N 71

a bit more easily understood. The new study on which Secretary Califano based his 20 percent estimate was written by 10 Public Health Service scientists, including the director of the National Cancer Institute (13). They were quite explicit in their criticism of the earlier -- and lower -- estimates of work- related cancer by half a dozen epidemiologists here and abroad, all of whom, interestingly enough, are longtime foes of ciga- rettes and some of whom have attributed a specific percent of cancer deaths to cigarette smoking. The PHS scientists wrote that "it is a reductionist error and not in keeping with current theories of cancer causation to attempt to assign each cancer to an exclusive single cause...Patterns and trends in total cancer incidence (and mortality) in the U.S. are consistent with the hypothesis that occupationally-related cancers comprise a substantial and increasing fraction of total cancer incidence." "The fact is that cancer is a disease of interac- tions," one of the authors said subsequently. "By ascribing a cancer to a single cause, we precluded looking at other causes" (14). Another of the 10, who described himself as "no fan of cigarette smoking," said that "you can't lay [all] the blame [on cigarettes]. The causes of cancer are much more complex. It is caused by the interaction of many factors" (15). 72

The new study originated during informal discussions by scientists who were skeptical about low occupational cancer estimates (16). The 10 scientists completed it for submission to the Occupational Safety and Health Administration (OSHA) during hearings on sweeping new rules proposed by OSHA to, regulate cancer-causing substances in the work place (17). Secretary Califano said that the 10 authors of the study, who included physicians and biostatisticians, actually estimated that between 21 and 38 percent of all cancer cases could be attributed to work exposure. But, he said, "we chose the more conservative figure because this is the first study of this kind we've done and because we'll be doing a lot more studies to refine it" (18). The PHS scientists attributed a sizable portion of occupational cancers to asbestos exposure (an estimated 67,000 deaths yearly), the rest to other substances associated in past studies with increased cancer incidence or death rates (see Table 1). They listed a dozen occupations in which cancer deaths are abnormally high, but for which specific causes have been identified (see Table 2). no "Perhaps the most important lesson to be learned from the asbestos story," they wrote, "is that a major public health disaster can develop while its early manifestations are lost by being attributed to other factors" (13). 73

Table 1 Chemicals Associated with Cancer Induction in Man Chemical Substances Asbestos Arsenic Benzene Chromium Coal tar Affected Tissues an.d Organs Lung, pleura & peritoneum, esophagus, stomach, colon/rectum Respiratory tract Blood-forming organs Respiratory tract p it ch, volatiles & coke oven Lung, larynx, emissions skin, scrotum Iron oxide Lung, larynx Nickel oxides Petroleum distillates Respiratory tract Lung, larynx Source: Bridbord et al 1978 (13) Estimated No. Workers Potentially Exposed Estimated No. Excess Cancers Per Year 8,000,000 to 11,000,000 67,000 1,500,000 2,100 - 7,300 2,000,000 240 - 1,400 1,500,000 2,400 - 46,000 60,000 160 - 800 1,600,000 1,300 - 5,000 1,370,000 3,800 - 5,000 3,000,000 2,400 - 12,000 74

"The Asbestos Story" Claims are being made that smoking workers who are exposed to asbestos have significantly increased risks of developing lung cancer. However, those who make such claims often overlook the extremely complex relationship between occupational exposures and the increased risk of certain diseases that have been discussed by various scientists and researchers, including those who prepared the new PHS report. Such claims, moreover, do not take into account the fact that in most studies of asbestos workers smoking habits have not been determined. In some others, the data have been incomplete. Claims that asbestos workers who smoke do have an increased risk of developing lung cancer were made in a 1968 report (19). That study, by Drs. Irving Selikoff and Jacob Churg of Mount Sinai School of Medicine and E. Cuyler Hammond of the American Cancer Society, was the first to take smoking into account. In that study, Selikoff and his colleagues calculated that asbestos workers who smoked had 92 times the risk of dying of lung cancer than workers who had neither smoked nor been exposed to asbestos. They made this calculation on the ~ basis of records from an insulation workers union, which o ~ ~ reportedly showed that only asbestos workers who smoked had W 75

Table 2 Occupational Groups in Which Excess Cancer Incidence Has Been Reported Without Identification of A Specific Etiologic Agent Occupational Groups Coal miners Chemists Foundry workers Textile workers Printing pressmen (newspaper) Metal miners Coke by-product workers Cadmium production workers Rubber industry Processing Tire building Tire curing Furniture workers Shoe workers Leather workers Percent Cancer Site Excess Reported Stomach 40 Pancreas, lymphoid tissue Lung Mouth and pharynx Mouth and pharynx 64 79 Lung Large intestine, pancreas Lung, prostate Stomach, blood-forming organs, Bladder, brain, Lung Nasal cavity and sinuses Nasal cavity and sinuses, blood-forming organs Bladder 200 181 312 135 248 80 140 88 90 61 300-400 700 100 150 PQ cn 0 Source: Bridbord et al 1978 (13) ~ ~ .A w ~ -o 76

died of lung cancer. Based on the reported association of smoking with lung cancer mortality and the association of asbestos exposure with the same disease, they concluded that the combination of smoking and asbestos exposure creates a risk that is larger than the expected combination of the two separate risks. This whole-is-greater-than-the-sum-of-the-parts concept is referred to in epidemiology as "synergism". In considering studies which have suggested such an interaction between asbestos and smoking, a 1978 PHS publication on asbestos mentioned the Selikoff study. The report said that the findings do suggest such an effect. But, the report added, "The relation of the statistical interaction to the pathogenesis of lung cancer is uncertain" (20). The risk of lung cancer in nonsmoking asbestos workers was discussed recently. Speaking at a conference on pollutants and high risk groups, OSHA's director of carcin- ogen identification and classification said last June that recent studies indicate a five- to 15-fold excess risk of lung cancer in asbestos workers who do not smoke (21). Any conclusion that it is the smoking which is responsible for the reported increased risk of lung cancer in the smoking asbestos worker is not justified on the basis tv c1i 0 ~ ~ of available evidence. -p w ~ 10 oa 77

Mapping New Clues Just two months after his announcement that 20 percent of U.S cancer cases probably originate in the work place (1), Secretary Califano began a $40 million program to find out more about the chemical hazards that Americans may be exposed to at work, at home, wherever they are. He said the "phenomenal technological advances" in this country have brought with them serious health hazards, "as in the case of asbestos," and directed four of his HEW agencies to work quickly "and with all available resources to identify and control the many toxic substances to which our citizens are exposed." He said relatively few of the more than 7 million chemicals now known to man have been tested for carcino- genicity and that as many as 60,000 are now believed to be or have been in commercial use, with 600 to 700 new ones entering commerce yearly. He ordered stepped-up animal testing at HEW and put in charge of the whole program one of the authors of his cancer-in-the-work-place report, Dr. David P. Rall, director of the National Institute of Environmental Health Sciences (22). The expanded testing program may provide new clues about those substances which apparently cause cancer in animals, and therefore may cause cancer in humans. This 78

information may assist epidemiologists in analyzing unusual geographic patterns they have found in cancer incidence. These patterns have been documented in several recent NCI studies including: • A compilation for every county in the contiguous 48 states of numbers of cancer deaths and age-adjusted death rates, by site, sex and race for the years 1950-69 (23). • The "Atlas of Cancer Mortality for'U.S. Counties 1950-69," w hich gives the same information graphically, with color-coded maps showing geographic variations in cancer incidence at the county level. The authors noted that "the maps for lung cancer indicate that excessive mortality is not limited to highly populated urban areas where cigarette smoking and air pollution are most prominent" (emphasis added). The maps enable epidemiologists in both public and private sectors to do correlation studies with data on demographic and environ- mental variables collected on a county level (24). • A similar atlas with data on U.S. nonwhites (25). • Another set of maps indicating counties in which 18 dif- ferent manufacturing industries have facilities and associated demographic data concerning employment and other population information. These maps, the authors wrote, reveal industrial county clusters that may provide a "useful point of departure" N) cn 4 ~ ~ ~ w ~ 0 0 79

for more in-depth local studies to evaluate occupational factors in cancer (26). The cancer atlases have been the basis for many studies already. NCI researchers, for instance, recently compared cancer mortality rates for counties with and without petroleum refineries and discovered an excess of cancers of the lung, nasal cavity and sinuses in those counties with refineries. They drew no firm conclusions about cause, but they surmised that the high rates of lung cancer in women might be due to a"pollution hazard spreading beyond the workplace". They did suggest industry-wide epidemiological studies to clarify the cancer risks among various groups of petroleum workers and to "evaluate the possible effects of petrochemical emissions released into neighboring commu- nities" (27). "There's something horrible out there, but we don't know what it is," one of the researchers told Business Week (28). The magazine reported in the same story that one large oil company had just added an epidemiologist to its staff and, with other companies, was participating in a tumor- registry program in which all cancer morbidity and mortality records of employees are funneled through the American Petro- leum Institute to Sloan-Kettering Institute for Cancer Research in New York. There, researchers look for patterns that might point to a cause. 80

Cancer epidemiology, incidentally, has been until recently the specialty of a relatively small number of sta- tisticians and health practitioners, whose work was largely sponsored by the government. The circle is tight and there is great demand on those who are already doing most of the work. Joseph F. Fraumeni, Jr., for instance, who co-authored the petrochemical survey (27), was also one of the prime movers behind the 1976 cancer atlas (24) and a member of the team of scientists who wrote Secretary Califano's cancer-3n- the-work-place report (13). Fraumeni told Business Week that NCI had doubled its epidemiology staff in two years and, like other scientists quoted in the article, bemoaned the lack of qualified environ- mental researchers to fill newly created positions in govern- ment and industry (29). Conclusion Seven years ago, Dr. Wilhelm Hueper -- one of the deans of research on occupational carcinogenesis -- warned that: Human exposure to many of these agents [potential human carcinogens] is not only widespread, but also often intense, and most of them are used without observing any real precautions. This lack of precau- tions may continue as long as the exposed public can be persuaded that the main lung cancer hazards are limited to cigarette smoking (30). 81

Dr. Hueper, whose occupational research on cancer began in the 1920s when he noted an alarming increase. in lung cancer since the turn of the century around industrialized cities in Central Europe, received a special Public Health Service award last fall mere months before his death at 84, for his "vision and courage" (31). The new PHS report may well focus needed atten- tion -- as recognized so long ago by Dr. Hueper -- on the role of occupational exposures in the development of disease. 82

References for Cancer in the Work Place 1. Califano, J.A. Remarks of the Secretary of Health, Education and Welfare to the AFL-CIO National Conference on Occupational Safety and Health, Washington, D.C., Sept. 11, 1978. 2. Medical World News. HEW Study Blames Bigger Fraction of Cancers on Jobs. Oct. 2, 1978, pp. 18-19. 3. Gehring, P. Study Sees 20% of Cancer Cases As Work- Related. The Washington Post, Sept. 12, 1978, pp. Al, A6. 4. Lang, R.A. Letter: The Washington Post, Sept. 26, 1978. 5. Roland, R.A. News release from the Manufacturing Chemists Association, Washington, D.C., Sept. 13, 1978. 6. Whelan, E.M. News release from the American Council on Science and Health, New York, Sept. 13, 1978. 7. Harding, R.L. , Jr. News release from the Society of the Plastics Industry, Washington, D.C., Sept. 15, 1978. 8. Hoerger, F. In: United Press International wire dis- patch, 7:31 PED, Oct. 25, 1978. 9. Leffall, L.D. In: Some Scientists Doubt New Cancer Estimates. The Tobacco Observer 3/5: October 1978. 10. Peto, R. In: Some Scientists Doubt New Cancer Estimates. The Tobacco Observer 3/5: October 1978. 11. Gori, G.B. In: HEW Study Blames Bigger Fraction of Cancers on Jobs. Medical World News, Oct. 2, 1978, p. 18. 12. Wynder, E.L. In: HEW Study Blames Bigger Fraction of Cancers on Jobs. Medical World News, Oct. 2, 1978, p. 18. 13. Bridbord, K., et al. Estimates of the Fraction of Cancer in the United States Related to Occupational Factors. National Cancer Institute, National Institute of Environ- mental Health Sciences, National Institute for Occupa- tional Safety and Health, U.S. Department of Health, Education and Welfare. Washington, Sept. 15, 1978. N Cn 0 ~ -~ ~ w ~ 0 P I b 83

14. Schneiderman, M.A. In: Increase in Estimate of Work Related Cancer Due to Change in Study Method. The Cancer Letter 4/40: 6-7, Oct. 6, 1978. 15. Bridbord, K. In: Govt. Scientists Discuss New Theory. The Tobacco Observer 3/5: 1, October 1978. 16. Saffiotti, U. Interview with The Tobacco Observer during the National Institutes of Health Conference on Cancer Prevention -- Quantitative Aspects, Reston, Va., Sept. 25-28, 1978. 17. News release from the Occupational Safety and Health Administration, U.S. Department of Labor, May 15, 1978. 18. Califano, J.A. Interview on CBS Morning News, Sept. 12, 1978. 19. Selikoff , I.J., et al. Asbestos Exposure, Smoking, and Neoplasia. Journal of the American Medical Associa- tion 204/2: 106-112, April 8, 1968. 20. U.S. Public Health Service. Asbestos: An Information Resource. National Cancer Institute, U.S. Department of Health, Education and Welfare. Bethesda, Md., DHEW Pub. No. (NIH) 78-1681, May 1978, 105 pp. 21. Infante, P.F. Occupational Health Practice and High Risk Groups. Presented at Conference on Pollutants and High Risk Groups, Amherst, Mass., June 5-6, 1978. 22. National Institutes of Health. Major Program Will Combat Threat of Chemical Hazards in Environment. NIH Record, Nov. 29, 1978. 23. Mason, T.J., and McKay, F.W. U.S. Cancer Mortality by County: 1950-1969. U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (NIH) 74-615, 1973, 729 pp. 24. Mason, T.J., et al. Atlas of Cancer Mortality for U.S. Counties: 1950-1969. National Cancer Institute, U.S Department of Health, Education and Welfare. Washington, DHEW Pub. No. (NIH) 75-780, 1976, 103 pp. 25. Mason, T.J., et al. Atlas of Cancer Mortality Among U.S. Nonwhites: 1950-1969. National Cancer Institute, U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (NIH) 76-1204, 1976, 142 pp. rv U7 O F_& 84

26. Stone, B.J., et al. Geographic Patterns of Industry in the United States: An Aid to the Study of Occupational Disease. Journal of Occupational Medicine 20/7: July 1978. 472-477, 27. Blot, W.J., et al. Cancer Mortality in U.S. Counties with Petroleum Industries. Science 198: 51-53, Oct. 7, 1977. 28. Blot, W.J. In: Using Cancer's Rates to Track Its Cause. Business Week, Nov. 14, 1977, pp. 69-70, 75. 29. Fraumeni, J.F., Jr. In: Using Cancer's Rates to Track Its Cause. Business Week, Nov. 14, 1977, pp. 69-70, 75. 30. Hueper, W.C. Lung Cancer and Smoking in Perspective. In: Frankel, C.J., editor. Lawyers' Medical Cyclopedia of Personal Injuries and Allied Specialties. Vol. 5 Revised, Pt. B, Sections 37.la Co., Indianapolis, 1972. to 38.99. The Allen Smith 31. National Institutes of Health. Hueper Receives Direc- tor's Award. NIH Record, Oct. 3, 1978. 85

Lung Cancer Carcinoma of the lung was one of the first diseases to be associated, in modern times, with tobacco smoke (1). Some government officials and scientists opposed to smoking' explain the reported dramatic increase in lung cancer death rates observed in this century by pointing to the concurrent rising popularity of smoking. They have also tried to attri- bute the apparent recent decline in the rate of increase of male lung cancer mortality to a decrease in smoking and the introduction of the new low "tar" cigarettes. These may be The failure to consider critically 1) important diagnostic advances, 2) changes in the reported frequencies of lung cancer cell types and 3) trends in cigarette consumption and lung cancer mortality data raises serious questions about any conclusions regarding smoking. easy and elementary explanations of lung cancer causation, but they seriously oversimplify the situation and ignore critical questions which need to be examined. Diagnostic Error A critical examination of the reported increase in rv lung cancer mortality must take into account changes in o ~ diagnostic techniques. The increase, claimed by some to be $ w ~ 0 v 87

"epidemic," may in fact have been created largely by radical improvements in diagnostic techniques that have become avail- able to physicians. In other words, more lung cancer has been found because physicians were better equipped to find it. Ramifications of this new ability to find lung cancer were discussed more than 20 years ago by a National Cancer Institute biostatistician. In 1955, Alexander H. Gilliam figured out that if the error in diagnosis of lung cancer were only 10 percent in 1914, decreasing to 2 percent in 1950 After introduction of most of the new diagnostic tools and techniques we know today, male lung cancer death rates would have only doubled in the 35-year period instead of the 26-fold increase indicated by national records (2). Further reason to question the validity of the "epidemic" can be found in what has been termed the tendency toward overdiagnosis of lung cancer. It has been suggested that primary lung cancer has become a "popular" disease to diagnose in smokers -- in other words, physicians are finding more lung cancer (whether or not it actually exists) because they are looking more for it. One New York chest specialist put it this way: The prodigious increase in lung cancer during the past three decades is not due to the exposure of the population to an alleged carcinogen but is the natural consequence of the widespread use of tech- niques not previously available. The intense interest in lung cancer has also produced a tendency 88

toward overdiagnosis of the disease on the basis of radiologic, biopsy, and cytologic findings which are often not substantiated by autopsy (3). Trends in Death Rates and Smoking Patterns Even if one assumes for the sake of argument that at least a portion of the "epidemic" is real, the trends in lung cancer death rates cannot be explained satisfactorily by smoking patterns. As long as 20 years ago, statisticians forecast a decline in lung cancer mortality. They based their predictions primarily on analyses of lung cancer death rate trends of the first half of the century. They also acknowledged such factors as the dramatic increase in the segment of the population most susceptible to cancer (the aged) and the overdiagnosis of the disease in smokers with accompany- ing underdiagnosis in nonsmokers (4-6). In 1961, the NCI statistician, Dr. Gilliam, made a forecast about U.S. lung cancer death rates, basing it on his analysis of mortality trends. A decline since 1948 in the rate of increase in lung cancer mortality, he said, indicated that "the disease will reach a peak among the white male population in the foreseeable future and then start to decline" (7). Indeed, within six months a researcher reported 89

that lung cancer death rates in the Seattle area had ceased to rise in persons under 60 and predicted that they would plateau in 10 to 15 years (8). In 1965, Great Britain's Registrar General made a similar forecast for his country (9), as did Canada's director of cancer statistics in 1966 (10). Subsequently, NCI statisticians attempted to equate the leveling lung cancer death rates in men with changes in their smoking rates (11). And others 3umped on the bandwagon, also attributing the change in mortality rates to smoking trends (12, 13). Another epidemiologist was more skeptical, however, when he wrote in 1974 that it did not seem likely that smoking patterns could explain lung cancer in the U. S., England and Wales. Ian Higgins, who is well-known for his anti-tobacco views, said that the flattening/decline of age-specific mor- tality rates in British men "appears to have preceded the reduction in cigarette smoking" (14). He even suggested that Britain may be "witnessing a saturation phenomenon. . .that those most susceptible to the disease have now developed it, with the result that the peak of the epidemic is now past"! Further support for rejecting the smoking-causation interpretation was provided in 1975 by a British thoracic surgeon, Dr. J. R. Belcher, who noted the changes in age 90

and incidence patterns of lung cancer in his country. He speculated on what caused the changes: Are they due to the discovery of the relationship of cigarette smoking to bronchial carcinoma and the -subsequent national campaign against the habit? This seems a likely suggestion until it is realized that the fall in the percentage increase in the rate and eventually of the rate itself in the younger age groups was happening as long ago as 1950. It seems more likely that the fall in the percentage rate of increase which dates back for at least fifty years has eventually led to an actual fall in the rate itself. This process has progressed steadily over many years, and represents the natural history of carcinoma of the bronchus (15). Even more recently, Professor Phillip Burch of the University of Leeds summarized much of his own work published since 1972 demonstrating the lack of correlation between smoking trends and changes in British lung cancer death rates. In January 1977 Burch wrote in the British Medical Journal: [T]he detailed changes in recorded death rates from lung cancer in England and Wales from 1901 to 1970 were strikingly synchronous in the two sexes. Thus the major cause of the increases had a simulta- neous impact on both sexes and could not have been cigarette smoking because the increase in consumption of cigarettes by women lagged some 30 years behind that of men (16). Burch said that post-mortem studies of the frequency of lung cancer indicated that the most important factor in the increase of recorded lung cancer has been clinical diagnostic error. 91

Self-Selection -- The Constitutional Hypothesis In 1976, the ACS interpreted data from its 25-state study as indicating that men who smoked cigarettes with lower "tar" and nicotine yields had lower lung cancer mortality (17). Commenting on these data in 1977, however, one British researcher wasn't so sure. Dr. M.A.H. Russell, who has long had strong anti-smoking views, said that drawing such a conclu- sion at the present time would be premature, "as the smokers who changed their cigarettes were self-selected" (18). Self-selection is an important part of the "constitu- tional hypothesis" first injected into the smoking and health controversy more than 20 years ago by Sir Ronald Fisher, recognized then -- as now -- as the father of modern-day statistics (19). The Surgeon General's report in 1964 des- cribed it as an "alternative hypothesi s that both the smoking of cigarettes and cancer of the lung have a common cause which determines both that an individual shall become a smoker and also that he shall be predisposed to lung cancer" (1). The constitutional hypothesis has been discussed frequently in recent years. Professor Burch in England, for instance, has maintained that the data on smoking and mortality in his country are more consistent with the constitu- tional theory than the smoking-causation theory. And he hasn't hesitated to say so. 92

Since 1972 Professor Burch has been published exten- sively on his theory and has provoked considerable discussion and controversy in the medical literature. But none of this dialogue within the scientific community about the cause of lung cancer has appeared in any of the HEW reports to Congress on smoking. Different Types of Lung Cancer That there are several distinct types of lung cancer -- distinguished by the appearances of the cells in the tumor under microscopic examination -- is well-known. Some types have stronger statistical associations with ciga- rette smoking than other types. Epidemiological studies have indicated, for instance, that cigarette smoking is more strongly associated with epider- moid, or squamous-cell, cancer of the lung and is not associ- ated, or only weakly so, with another major type, adeno- carcinoma. There has been little discussion of lung cancer by cell type in the yearly HEW reports to Congress on smoking and health, but it has been generally accepted that adeno- carcinoma is more common in women and in nonsmokers. In 1977, a New York researcher, Dr. Ronald G. Vincent, reported that pathologists at Roswell Park Memorial Institute were finding a rapid decrease in squamous-cell carcinoma, and 93

Figure 1 The Changing Histopathology of Lung Cancer 1963-1975 N =1682 ___-. -Adeno // / - Squamous i . -- Smal I Large Undiff Broncheo- '•`• ~Aiveolar ~ ~'~..~~....~•••.~..~~....~..a •• ~.. Mixed V Source: Vincent et ai.1977 (20). 94

a corresponding increase in adenocarcinoma (see Figure 1). He added that if their data proved to be representative of national trends, "adenocarcinoma will soon become the most prevalent type of lung cancer in the United States" (20). Speculating about the factors responsible for this development, Dr. Vincent mentioned the increasing incidence of lung cancer in women, modifications in the way pathologists identify lung cancer cell types and environmental and occupa- tional agents and alterations that have occurred over the last few years in cigarettes. He conceded, however, that his team had considered such factors as "length of smoking history, form of tobacco used, quantity of tobacco used, age habit started, degree of inhalation and the use of filters", and that they had been "unable to equate the histology of lung cancer with any of these factors". The reasons for this development, if it ultimately proves to be representative of the national experience, remain unknown. One wonders if this reported increase in adeno- carcinoma may have a similar basis as the over-all lung cancer "epidemic" discussed above -- that is, observer variations. Much of the over-all "epidemic" may be due to changes in clinicians' techniques; much of the adenocarcinoma increase may be due to modifications in the way pathologists classify lung cancer cell types. 95

As a case in point, three pathologists in Connecticut reappraised the histopathology of 449 lung cancer cases from original tissue specimens and showed the rate of agreement with past readings was only 63 percent (21). They reviewed cases that had been interpreted between 1953 and 1959. On their reappraisal, the percentage of cases classified as adenocarcinoma increased markedly, while the percentage of squamous-cell carcinoma decreased significantly. Those investigators suggested that a reappraisal of associations between specific histological types of lung cancer and smoking was warranted. New Hypothesis -- Diet Other developments in lung cancer in recent years include a new emphasis on what might be called the "diet hypothesis". An epidemiologist in Japan who is known to be strongly anti-tobacco reported last September that eating green and yellow vegetables daily lowered the risk of lung cancer in smokers and nonsmokers alike (22). "This came as a surprise to me, because I thought that only cigarette smoking could influence the risk of lung cancer," Takeshi Hirayama told a Medical World News reporter (23). Studies in animals and humans in recent years have indicated that dietary vitamin A, which green and yellow 96

vegetables are high in, might have an inhibitory effect on pulmonary carcinogenesis. For example, a study in Norway considered vitamin A intake and smoking habits in relation to subsequent development of lung cancer. The author said that his findings were in accord with experimental result•s in animal studies and "call for further exploration of the role of nutritional factors in the development of lung cancer" (24). He added that the association of lung cancer with cigarette smoking "may have held back" attempts to study relationships between nutritional factors and lung cancer. Whether regular consumption of green and yellow vegetables or a deficiency of Vitamin A should be considered important in lung cancer causation is not clear. But a special task force meeting in Stockholm's famous Karolinska Institute in 1977 took it seriously. Although the scientists were strongly opposed to smoking, they said that the evidence on vitamin A is "thought-provoking and needs to be followed up" (25). Animal Experimentation Results of some animal experiments have been applied to man to support the claim that cigarette smoke causes cancer. There are many serious, perhaps irresolvable, pro- b lems in any extrapolation of animal results to humans. Lung 97

cancer of the type associated with cigarette smoking in men has not been produced in animals in inhalation studies (26). And severe criticism has been levelled at certain of the studies because of unrealistic experimental situations which could scarcely be comparable to the human experience -- for instance, extremely high dosages over a short period of time (27-29). There are, obviously, also significant differences between the tissues of man and laboratory animals, and that makes it difficult to draw any conclusions about the human relevance of animal results. Conclusion The claim that cigarette smoking causes lung cancer has not been scientifically proven. The charge ignores basic unresolved scientific questions concerning cell types, animal experimentation, smoking patterns and lung cancer rates, dietary influence and diagnostic variations. Lung cancer is a complex disease, and a one-sided attack on cigarette smoking as the causal agent does nothing to advance the search for its cause and cure. Within a generally negative presentation on smoking, the pathology department chairman at UCLA told a Public Health Service meeting only a year and a half ago: Although epidemiological data has clearly established cNj+ the existence of a correlation between smoking and ~ 98

[lung cancer], a clear-cut causal relationship between cigarette smoking and cancer has not been demonstrated (30). 99

References for Lung Cancer 1. U.S. Public Health Service. Smoking and Health. Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington, U.S. Department of Health, Education and Welfare, PHS Pub. No. 1103, 1964, 387 pp. 2. Gilliam, A.G. Trends of Mortality Attributed to Carcinoma of the Lung: Possible Effects of Faulty Certification of Deaths to Other Respiratory Diseases. Cancer 8/6: 1130-1136, 1955. 3. Rosenblatt, M.B. The Increase in Lung Cancer: Epidemic or Artifact? Medical Counterpoint: 29-40, March 1969. 4. Lees, T.W. Smoking and Lung Cancer. The Darien Press, Ltd., Edinburgh, 1959, 32 pp. 5. Lees, T.W. Letter: Lancet 1: 1116-1117, 1965. 6. Lees, T.W. Letter: Lancet 2: 443, 1965. 7. Gilliam, A.G., et al. Trends of Mortality Attributed to Carcinoma of the Lung: The Declining Rate of Increase. Cancer 14/3: 622-628, May-June 1961. 8. Ravenholt, R.T. In: Lung Cancer, Heart Ills May Be Near Peak, by John Barbour. Louisville Times, Nov. 13, 1961. 9. General Register Office. The Registrar General's Statis- tical Review of England and Wales for the Year 1962: Part III, Commentary. London, Her Majesty's Stationery Office, 1965. 10. Phillips, A.J. An Analysis of the Increase in Lung Cancer in Canada. Canadian Medical Association Journal 95: 1172-1174, Dec. 3, 1966. 11. Schneiderman, M.A. and Levin, D.L. Trends in Lung Cancer: Mortality, Incidence, Diagnosis, Treatment, Smoking and Urbanization. Cancer 30/5: 1320-1325, November 1972. 12. Wynder, E.L. Etiology of Lung Cancer: Reflections on Two Decades of Research. Cancer 30/5: 1332-1339, November 1972. 13. Horn, D. In: Lung Cancer Dropping Among American Men, by Joann Rodgers. Baltimore News American, Oct. 24, 100

1974, p. 3A. 14. Higgins, I.T.T. Trends in Respiratory Cancer Mortality in the United States and in England and Wales. Archives of Environmental Health 28/3: 121-129, March 1974. 15. Belcher, J.R. The Changing Pattern of Bronchial Carcin- oma. British Journal of Diseases of the Chest 69: 247-258, 1975. 16. Burch, P.R.J. Letter: British Medical Journal 1: 165, 1977. 17. Hammond, E.C., et al. "Tar" and Nicotine Content of Cigarette Smoke in Relation to Death Rates. Environ- mental Research 12: 263-274, 1976. 18. Russell, M.A.H. Smoking Problems: An Overview. In: Jarvik, M., et al, editors. Research on Smoking Behavior. National Institute on Drug Abuse. U.S. Department of Health, Education and Welfare. Rockville, Md., NIDA Research Monograph 17, December 1977, pp. 13-34. 19. Fisher, R.A. Smoking: The Cancer Controversy, Some Attempts to Assess the Evidence. Oliver and Boyd, Edinburgh and London, 1959, 47 pp. 20. Vincent, R.G., et al. The Changing Histopathology of Lung Cancer: A Review of 1682 Cases. Cancer 39: 1647-1655, April 1977. 21. Yesner, R., et al. A Reappraisal of Histopathology in Lung Cancer and Correlation of Cell Types with Antecedent Cigarette Smoking. American Review of Respiratory Disease 107: 790-797, 1973. 22. Hirayama, T. Diet and Cancer. Delivered at National Cancer Institute Conference on Cancer Prevention - Quan- titative Aspects, Reston, Va., Sept. 25-28, 1978. 23. Hirayama, T. In: Japan: Colon Ca Moving Up. Medical World News, Nov. 13, 1978. 24. Bjelke, E. Dietary Vitamin A and Human Lung Cancer. International Journal of Cancer 15/4: 561-565, 1975. 25. Cederlof, R., et al. Air Pollution and Cancer: Risk Assessment Methodology and Epidemiological Evidence: Report of a Task Group. Environmental Health Perspec- tives 22: 1-12, February 1978. 26. Furst, A. Statement for the Subcommittee on Health, 101 '_&

Committee on Labor and Public Welfare, U.S. Senate, 94th Congress, Second Session, Cigarette Smoking and Di- sease, 1976, Feb. 19, 1976: pp. 100-116. 27. Kessler, I.I. Saccharin, Cyclamate, and Human Bladder Cancer: No Evidence of an Association. Journal of the American Medical Association 240/4: 349-355, July 28, 1978. 28. Stauffer, H.P., and Riedwyl, H. Interaction and pH 9. Dependence of Effects of Nicotine and Carbon Monoxide in Cigarette Smoke Inhalation Experiments with Rats. Agents and Actions 7/5-6: 579-588, 1977. Bair, W.J., et al. Apparatus for Direct Inhalation of Cigarette Smoke by Dogs. Journal of Applied Physiology 26/6: 847-850, 1969. 30. Van Lancker, J.L. Smoking and Disease. • In: Jarvik, M.E., et al, editors. Research on Smoking Behavior. Jarvik, M.E., et al, editors. National Institute on Drug Abuse. U. S. Department of Health, Education and Welfare. Rockville, Md., NIDA Research Monograph 17, December 1977, pp. 230-280. 102

Other Cancers HEW's yearly reports to Congress have cited various prospective and retrospective population surveys to support claims of statistical relationships between cigarette smoking and cancers of the oral cavity, bladder, esophagus, larynx and pancreas (1, 2). However, close examination of the inci- dence rates of these diseases reveals many unexplainable patterns that are almost impossible to reconcile with the hypothesis that cigarette smoking causes them. The incidence rates usually cited for these cancers, and their time patterns, have been derived from data collected The establishment of any relationship between smok- king and cancers of the larynx, esophagus and bladder must involve considerable guesswork, because of the vastly different incidence patterns and trends of these diseases and multiple suspected causes. in two large-scale studies, conducted 25 years apart, by the Public Health Service: The Second National Cancer Survey (SNCS), 1947-49, and the Third National Cancer Survey (TNCS), 1969-71 . Source data from the First National Cancer Survey, 1937-39 , are no longer available, but comparative data from the Second and Third, for the seven geographic areas that they had in common, were published by the National Cancer ~ 0 ~ Institute last year (3). More than 20,000 cancer cases were ~ included in the SNCS, more than 125,000 in the TNCS. . to N N w 103 ,

Figure 1 Incidence Rates* per 100,000 Population from Second National Cancer Survey (SNCS) (1947-49) and Third National Cancer Survey (TNCS) (1969-71) SNCS 24 21 -I- 18 + 15 -I- 12 4- 9+ 6 -E- 3 -F- TNCS SNCS 0 0 6 0, CANCER ORAL CAVITY SITE & PHARYNX I BLADDER Q- white male L - nonwhite male ESOPHAGUS [:] - white female 0- nonwhite female *Age adjusted to the 1950 U.S. population standard. Source: Devesa and Silverman 1978 (3). TNCS SNCS TNCS N Cn O H ~ -Q' W N N 4b 104

The graphs in Figure 1 illustrate such incongruities as incidence rates rising, falling or remaining stable depend- ing on disease, gender and race. Even when one considers the so-called "lag period" -- the time between the action of the alleged causal agent (starting smoking) and the diagnosis of the disease -- the graphs still do not support the claim that cigarette smoking is the causal agent. For example, nonwhite male rates for cancers of the oral cavity and pharynx were up by almost 40 percent in 25 years, while the rates for white men and women were down by approximately 30 percent. The rates for nonwhite females decreased, too, but not so sharply. Further, similar peculiarities in the rate patterns can be observed for cancer of the bladder. The rates for males, both white and nonwhite, increased markedly -- especially the nonwhite male rate, which nearly doubled. The rates for both groups of females were sharply down. Some critics of cigarettes have claimed that blacks and other nonwhites may have been slower to take up cigarette smoking in any numbers and are therefore later in showing increases in the "cigarette-related" diseases. This contention is not supported by the data, as illustrated in Figure 1. On the other hand, the National Cancer Institute researchers have suggested another explanation for these trends in nonwhites, namely that other factors such as access to medical care and improved diagnosis may partially explain the increased inci- dence in some of these diseases in blacks. 105

But these factors can hardly account for the Jump in esophageal cancer in nonwhite males while white male incidence was dropping, the increase in bladder cancer in nonwhite men while it dropped in nonwhite women or the decrease in oral/pharyngeal cancer in nonwhite women while it climbed steeply in nonwhite men. Other published studies and surveys indicate that no firm statements of causality correctly can be made about smoking and any of these cancers which tically linked with cigarettes. Oral/Pharyngeal Cancers have been statis- Although claims have been made that smoking causes oral and pharyngeal cancers, such contentions are not supported by an examination of the available literature. A number of studies, for example, have failed to establish a relationship between smoking and oral cancer (4, 5) and smoking and pharyn- geal cancer (5). Recent studies on tobacco and oral and pharyngeal cancers have examined the possibility of a substantially higher risk for development of these cancers in individuals who both smoke and drink. One study in 1977 reported such an increased risk in patients examined in 20 hospitals in eight American cities (6). However, this was not confirmed by two ' 1. 106

other studies of oral cancer patients published the same year (7, 8). Interestingly enough, one (7) pointed to another factor -- poor dentition or tooth spacing -- as a more impor- tant risk factor than either cigarettes or alcohol. Attempts to study the possible influences of ciga- rettes and alcohol in the development of oral and pharyngeal cancers are complicated, because individuals who smoke also are more likely to drink. Therefore, it is difficult to determine what -- if any -- roles these social habits play. Esophageal Cancer A biostatistician from the International Agency for Research on Cancer in France recently reviewed worldwide patterns of esophageal cancer and certain epidemiological data (9). He concluded that "the data strongly suggest that factors associated with poverty and specific limitations of dietary intake increase susceptibility for this disease." This researcher noted "exceptionally high rates in some areas," and "a higher incidence among the lower socio- economic groups," most markedly in females. He said that although there is a statistical association with tobacco and heavy alcohol use in the U.S. and Western Europe, in much of the rest of the world they are not "factors of major impor- tance." 107

"A considerable array of external factors have been associated with the disease," he said, "but none strongly enough or with sufficient consistency between countries for their etiological role to be basic." In another recent study, researchers examining the incidence of esophageal cancer in Southern Iran found that the risk of this cancer in males and females was "at least nearly equal," even though cigarette smoking and alcohol consumption are "almost exclusively male" habits in this region (10). This observation led them to speculate that "it seems probable that the etiology of the disease in this region is not explainable by association with tobacco smoking and/or alcohol use, since these habits do not correspond with the epidemiology of the disease as found in this study." Laryngeal Cancer After examining nearly 200 patients with laryngeal carcinoma, an English otolaryngologist questioned the claim that smoking causes this cancer (11). He observed that "the incidence of laryngeal cancer has remained more or less constant for 70 years -- a period in which tobacco consumption (has] risen sharply." He also warned that "any data showing a correlation between heavy cigarette smoking and ~ 0 laryngeal carcinoma must be interpreted with caution because p .a in the entire population the incidence of laryngeal carcinoma ~ m 108

has been remarkably constant." The British physician's conclusions are supported by the work of an Argentinian who examined the smoking habits of 187 cancer patients in three cities (12). Statistical analysis of the data, he said, revealed "values [for smoking] that are not significant, consequently not indicating a depen- dency between the presence of the studied cancer and the smoking habit." He said, "We can therefore hypothetically assume that other factors, besides the significance of the smoking habit, must logically affect the etiology of these pathologies." Smoking apparently also was discounted as a causal factor by a British scientist who investigated the possible relationship between alcohol and tobacco use and laryngeal cancer by examining disease and death rate patterns (13). One problem, he said, is that trends in mortality rates in different age groups show contrasting characteristics. He found, after drawing up time trend charts by five-year age groups, that the more or less consistent fall in death-rates in both sexes -- while per capita consumption of alcohol and cigarette use rose in both sexes -- "would seem to be incompatible with the hypothesis that tobacco and/or alcohol are major causal agents." N Before his discussion of the data on laryngeal o N -pb cancer, he had commented on the need to use all available data 4' 109 -

in attempting to determine disease causation: "There can be .I- no doubt that epidemiologists have the professional respons- ibility for elucidating the causes of disease with all the ingenuity and thoroughness they can command." But with a wryness rare in medical journals, he also noted that "crying wolf in the absence of the marauder is generally held to be counterproductive." Bladder Cancer A number of studies examining the incidence of bladder cancer have found no association between cigarette smoking and the occurrence of bladder tumors (14-16). Incon- sistencies in the conclusions of reports examining the rela- tionship between smoking and bladder cancer and "the relative weakness of the evidence for an association in females" led the journal Lancet to "suggest the need for caution in interpreta- tion" of such studies (17). A major paper on bladder cancer was published by the National Cancer Institute in late 1978 (18). The two government epidemiologists looked at the geographic patterns identified in their institute's "cancer atlas" (19) (see Cancer In Workplace chapter). Bladder cancer maps, according to these researchers, showed "significantly higher" rates among males in the North- 110

Figure 2 JNCI Journal of the National Cancer Institute October 1978 Volume 61 Number 4 N cn 0 ~ .p -a CA) r") CA) ~-x U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE Public Heafth Service National Institutes of Health 111

® east and in counties with heavy concentrations of plants manufacturing dyes and pigments, pharmaceutical preparations, perfumes, cosmetics and certain other toiletries. The authors confirmed an increase in bladder cancer in urban areas, which one of them had previously hypothesized to be due to heavier cigarette smoking in city dwellers (20). Now they said, "Cigarette smoking, however, is not likely to be responsible for the elevated mortality in the Northeast, in as much as the available national surveys show only small regional differences in smoking practices." They also noted that bladder cancer is the only "smoking-related" cancer for which death rates among males in Northern urban areas are lower in blacks than in whites. They surmised that the lower risk might be due to limited employment opportunities for blacks, especially pre-1960, in industries where workers may be exposed to chemical carcino- gens. In another study published in late 1978, NCI epidemi- ologists associated the levels of certain contaminants of municipal drinking water with bladder cancer in both sexes. The researchers said an interaction of chlorine with other substances in water treatment plants, and sometimes contamina- tion from industry create trihalomethanes (THMs) in varying quantities. The researchers said bladder cancer rates showed the strongest and most consistent association of all cancers 112

with their THM exposure index after control of other factors, including industrialization of the county studied (21). Cancer of the Pancreas The inconclusive nature of research in pancreatic cancer is illustrated by a statement in the 1976 report on "The Health Consequences of Smoking" (22). The authors of th e report examined the association with cigarette smoking and concluded that "the significance of the relationship is not clear at this time." The so-called "smoking-related" cancers have been reported to occur less frequently among predominantly Mormon populations (23). It has been hypothesized that the Mormons' abstemious way of life is responsible for the phenomenon, as they advocate no alcohol, tobacco, tea, coffee or drugs, especially the addictive sort, and they stress moderation in the use of meat. One exception to the Mormon hypothesis was demon- strated in a recent study in Utah that analyzed all cancer cases identified in the state between 1966 and 1970 and com- pared the incidence found in Utah residents (both Mormons and non-Mormons) to that of the total population covered in the Third National Cancer Survey (24). Records of more than 10,000 cases in the state 113

cancer registry showed no significant difference by religion in pancreatic cancer incidence. Both Mormons and non-Mormons, as a matter of fact, had a low incidence of the disease compared to TNCS incidence, which, said the Utah researchers, "raises the question of other, unidentified factors." Conclusion In a paper published in early 1978, one of the top epidemiologists at the National Cancer Institute reviewed, much as we did at the beginning of this chapter, the incidence of some of the cancers with which smoking has been linked (25). Discussing the trends in "head and neck cancers" he noted many of the same inconsistencies we did. And he said: For these diseases, at least, it is not likely that any one single etiology will give a satisfactory explanation...What does this imply? It implies that we've got to look for more clues. . . New clues should lead to new inquiries and new answers. 114

References for Other Cancers 1. U.S. Public of Smoking: and Welf are. 1974, 124 pp. 2. U.S. Public of Smoking: and Welfare. 1975, 235 pp. 3. Health Service. The Health Consequences 1974. U.S. Department of Health, Education Washington, DHEW Pub. No. (CDC) 74-8704, Health Service. The Health Consequences 1975. U.S. Department of Health, Education Washington, DHEW Pub. No. Devesa, S.S., and Mortality Journal of the March 1978. (CDC) 76-8704, and Silverman, D.T. Cancer Incidence Trends in the United States: 1935-74. National Cancer Institute 60/3: 545-571, 4. El-Mofty, S. Oral Cancer in the United Arab Republic. Oral Surgery, Oral Medicine & Oral Pathology 24/2: 240-245, August 1967. 5. Cederlof, R., et al. The Relationship of Smoking and Some Social Covariables to Mortality and Cancer Morbidity. A Ten Year Follow-Up of Smoking in a Probability Sample of 55,000 Swedish Subjects Age 18-69. Karolinska Insti- tute, Stockholm, Sweden, 1975, 91 pp. 6. Wynder, E.L., and Stellman, S.D. Comparative Epidemiology of Tobacco-Related Cancers. Cancer Research 37/12: 4608-4622, December 1977. 7. Graham, S., et al. Dentition, Diet, Tobacco, and Alcohol in the Epidemiology of Oral Cancer. Journal of the National Cancer Institute 59/6: 1611-1618, December 1977. 8. Browne, R.M., and Waterhouse, J.A.N. Alcohol and Tobacco Habits in Oral Squamous Cell Carcinoma. International Association for Dental Research Abstracts, No. 210, 1977. 9. Day, N.E. et al. Some Aspects of the Epidemiology of Esophageal Cancer. Cancer Research 35/11 (Part 2): 3304-3307, November 1975. 10. Sadeghi, A., et al. Cancer of the Esophagus in Southern Iran. Cancer 40/2: 841-845, 1977. 11. Stell, P.M. Smoking and Laryngeal Cancer. Lancet 1: 617-618, 1972. 115

12. Ferrara, F.A. Ecological Analysis of Lung Cancer in the City of La Plata. Proceedings of the 2nd International Clean Air Congress: 244-247,.1970. 13. Burch, P.R.J. Are 90% of Cancers Preventable? IRCS Medical Sciences 6: 353-356, 1978. 14. van der Werf-Messing, B., and Kaalen, J.G.A.H. Occupa- tions and Smoking Habits of Bladder Carcinoma Patients in Rotterdam. Jaarboek van Kankerondezoek en Kankerbest- rijding in Nederland No. 19: 77-85, 1969. 15. Anthony, H.M., and Thomas, G.M. Tumors of the Urinary Bladder: An Analysis of the Occupations of 1,030 Patients in Leeds, England. Journal of the National Cancer Institute 45/5: 879-895, November 1970. 16. Schievelbein, H., et al. Tryptophane Metabolism, Cancer of the Urinary Bladder and Smoking Habits. Zeitschrift fur Klinische Chemie und Klinische Biochemie 10/10: 445-449, 1972. 17. Editorial: Smoking and Cancers of the Bladder and Kidney. Lancet 1: 635-636, 1971. 18. Blot, W.J., and Fraumeni, J.F., Jr. Geographic Patterns of Bladder Cancer in the United States. Journal of the National Cancer Institute 61/4: 1017-1023, October 1978. 19. ton, DHEW Pub. No. (NIH) 75-780, 1976, 103 pp. Mason, T.J., et al. Atlas of Cancer Mortality for U.S. Counties: 1950-1969. National Cancer Institute, U.S. Department of Health, Education and Welfare. Washing- 20. Fraumeni, J.F., Jr. Cigarette Smoking and Cancers of the Urinary Tract: Geographic Variation in the United 21. 22. 23. 24. States. Journal of the National Cancer Institute 41/5: 1205-1211, November 1968. Cantor, K.P., et al. Associations of Cancer Mortality with Halomethanes in Drinking Water. Journal of the National Cancer Institute 61/4: 979-985, October 1978. U.S. Public Health Service. The Health Consequences of Smoking: A Reference Edition. U.S. Department of Health, Education and Welfare. Center for Disease Con- trol, Atlanta, 1976, 657 pp. ~_) cri 0 Enstrom, J.E. Cancer Mortality Among Mormons. Cancer ~ 36/3: 825-841 Se tember 1975. ~ , p ~ w Lyon, J.L., et al. Cancer Incidence in Mormons and N w 116

Non-Mormons in Utah, 1966-1970. New England Journal of Medicine 294/3: 129-133, Jan. 15, 1976. 25. Schneiderman, M.A. Time Trends: United States 1953- 1973. Laryngoscope Suppl. 88/1 Part 2: 44-49, January 1978. 117

Cardiovascular Disease In 1949 in Framingham, Mass., the U.S. Public Health Service started a program of close surveillance of more than 5,000 adult men and women selected by random sampling. The major objective was to attempt to determine why individuals would develop evidence of coronary heart disease (CHD). Through questionnaires and painstaking observations, the researchers would record the variables in lifestyle, environmental characteristics, familial traits and other factors believed to relate in any way to CHD. Then they would see which of these variables were most common in those persons A fair appraisal of the evidence, examined in its entirety, indicates that the risk of coronary heart disease is strongly associated with genetic and lifestyle factors. who did develop heart disease symptoms. And they would try to analyze statistically the relative importance of each of the variables in the occurrence of those symptoms. From that community-wide study, there developed the concept of "risk factors" in CHD and other diseases (1). The Framingham study originally found statistical relationships between heart disease and high serum cholesterol level, elevated blood pressure, cigarette smoking, obesity and low 119

vital capacity (2). The researchers stated last year that elevated blood pressure "has been confirmed as the dominant contributor" to heart disease in the Framingham study, but that they are continuing to study the possible role of other "risk factors" in the development of heart disease (3). The presence of an association, however, is neither proof of causality nor a demonstration that the elimination or reduction of a risk factor will prevent the occurrence of disease (4), though one might not know that from the statements of some government health officials (5-7). Dr. Christiaan Barnard, the famous heart surgeon, has provided an illustrative analogy on this point: [T]he statistician can do no more than point out the existence of an association between two variables; he cannot prove a cause-and-effect relationship between them. Stated more simply, an association between a particular diet and the incidence of coronary heart disease can be demonstrated, but that does not by any means Qrove that following that diet causes heart disease. I am sure an association could be shown between the absence of legs and the inability to respond to the verbal command "Jump." But this would not prove that people hear with their legs (8). The director of the Heart, Lung and Blood Institute reached such a conclusion, however, in addressing a meeting of medical writers in 1977. He told them that elimination of smoking would reduce CHD mortality by 150,000 deaths per year (9). Less than six months later, on Capitol Hill to justify his institute's budget, he told House Appropriations Committee 120 N U1 0 ~ -p -Ph. w N w ~O

members that "we still don't know the etiology of arterioscler- osis and hypertension" (10). He said his researchers are "testing the hyQothesis...that lowering cholesterol and cessation of smoking will delay or prevent the onset of heart disease" (emphasis added). It becomes obvious that scientists know little more about CHD causation now than they did in 1964 when the original Surgeon General's report said that "the basic cause or causes of coronary heart disease are obscure" (11). One thing scientists do know is that heart disease death rates are down. In fact, a meeting was called in Washington last October to discuss a 14-year low in heart disease death rates (12). Vital statistics presented at the meeting revealed a 24 percent decrease over-all in CHD since 1968, while deaths from all causes dropped only 17 percent. Rates for cerebrovascular diseases were down even more (13). See Table 1. Meeting participants, who included heart and public health specialists from across the nation, basically agreed that the decrease in the death rates for the nation's leading cause of death was real and not just some statistical aberra- tion. After all, the rates were down in both sexes, in both whites and nonwhites and in all age groups (14). But they could not decide whether improved treatment techniques of ~ O recent years or improvement in what heart researchers call ~ w N) 121 a

Table 1 Per Cent of Decrease in Mortality Rates United States, 1968 to 1976 Persons Age 35-74, by Sex and Race Cause of Death Per Cent of Decrease 1968 to 1976, by Sex and Race White White Nonwhite Nonwhite Men Women Men Women All Coronary Heart Disease -21.0% -26.5% -30.7% -39.1% -24.3% Cerebrovascular Disease -30.6% -30.4% -43.7% -47.1% -32.7% Major Cardiovas- cular Disease -20.9% -26.1% -33.2% -40.7% -24.6% All Causes -15.3% -16.4% -24.8% -32.7% -17.3% Source: Stamler 1978 (13) 122

"America's risk factor profile" should be given credit for the brighter picture (15). After all, as the Framingham project director had noted earlier in the year, many persons have become more health conscious and have "gradually changed their life-styles with regard to dairy products, exercise, weight-watching, antihyper- tensive treatment, and cigarette smoking." This has all coincided with the decline in coronary mortality, he said, "but the existence of a causal relationship is unclear" (16). Attempts to determine the causes of heart disease have been complicated in the last few years by the recognition of other possible risk factors. These include family history of heart attacks (17), urban vs. rural residence (18), viral infection (19) and a factor called "acculturation" (20, 21). This last variable -- to many, one of the more worthy of further investigation -- relates to the departure of younger persons from the traditional and family-oriented ways of their elders. As their lives change, so do their risks of major heart disease. And this enhanced risk can not be ex- plained statistically by changes in diet, cholesterol levels, blood pressure or smoking. A recent example of this was reported from a con- tinuing PHS-funded study of CHD incidence in Japanese citizens and individuals of Japanese ancestry in Hawaii and California 123

(20). Another example is the experience of a small, close-knit Italian village in Pennsylvania that over the years became more like a typical American suburb (21). The stresses of modern living extend from city to tiny mining community, from executive office to farmyard, which brings us to another relatively unexplored variable in heart disease, the Type A behavior pattern. Persons who exhibit a Type A behavior pattern are chronically in a pushing constantly to keep up with the Joneses, seeking hurry, recog- nition and advancement, suffering from what the two physician researchers who have identified the pattern call a "paucity of time itself" (22). These Type A persons, as one could guess, are more coronary prone. Four papers published since 1977 add to the growing literature on the subject and strengthen the belief of many scientists that further exploration is warranted (23-26). A potentially unifying link with a wide variety of these "risk factors" was postulated recently by a Columbia University physician researcher in a study of male heart patients. This researcher believes that an imbalance in sex hormones in the bloodstream may account for an individual getting heart disease (27). Commenting on his research, Dr. Gerald B. Phillips told The New York Times further studies are needed to examine 124

women and older men who have heart attacks. He said he expects to find the same imbalance in them. And if further studies by himself and others sustain his theory, he said, "it should not be difficult to change the blood hormone levels in order to prevent a heart attack"--- by diet, drugs or other means (28). The Times commented that if Dr. Phillips' theory is right, "hormone changes could explain why heart attacks occur most often in older men and in post-menopausal women" (29). Two British researchers also have speculated that hormones may play a role in the development of CHD (30). From their study of mortality rates in England and Wales, they concluded that "further studies are needed to clarify [the role of male sex hormones] in the aetiology of CHD in men." Despite this emphasis on the determination of "risk factors", physicians and researchers have questioned whether they are really relevant in any discussion of the development of CHD. For example, a West Virginia heart specialist wrote in American Heart Journal that "all of us know that every good cardiologist must repeat these words [he called hypercholesterolemia, smoking and obesity the ~ "magic incantation"] three or four times daily to reaffirm o F~ -P his belief in what has been accepted as causative factors ~ ~ 4~* ~ 125

of this dread illness. However, few of these relationships have proved definitive". He also suggested that: Our primary need in cardiology today is a few heretics who will abandon the practice of equating statistical association with the etiology of athero- sclerotic heart disease. This would mean a greater emphasis on the causation of atherosclerosis itself (31). A physician in Ireland considered what he called the "major risk factors", but his weren't exactly the same three as the West Virginian's. He came to the same general conclusion as the American, however, writing in Journal of the Irish Medical Association that there is "no evidence in man to support the contention that the control of elevated lipids, cessation of cigarette smoking or the control of hypertension retards the development of the atherosclerotic process or delays the rate at which coronary artery stenoses develop once these processes are established" (32). As a well-known American epidemiologist has written in New England Journal of Medicine limitations in the current knowledge of the etiology and methods of prevention of CHD "argue for broadening the search for contributing causes and possible dynamics of pathogenesis, rather than merely inten- sifying the study of the few traditional 'risk factors"' (33). 126

Carbon Monoxide Carbon monoxide (CO), a colorless, odorless and tasteless gas produced by the burning of any material contain- ing carbon, has been thought by many to be the component of cigarette smoke that might explain the statistical associa- tion between smoking and CHD. In early studies, rabbits chronically exposed to CO and fed a high cholesterol diet were found to have more arterial changes similar to early atherosclerosis in man than did animals on a similar diet but not CO-exposed (34-39). "Does CO play a role in arteriosclerosis?" one of the most prominent researchers in smoking and health asked rhetori- cally of an American Cancer Society audience in Philadelphia 18 months ago. Answering his own question, he said, "It certainly works in rabbits, but there's considerable doubt whether it works in man" (40). Well, now it doesn't appear to work in rabbits, either. Poul Astrup, one of the researchers who did the animal experiments with CO, recently reported that he and his group have been unable to reproduce their results (41-43). In a presentation describing these findings, they said that "no direct toxic effect of CO" could be demonstrated (42). 127

Nicotine Although claims have been made that nicotine causes heart disease, a number of researchers and physicians have questioned the basis for such assertions. For example, one heart specialist wrote recently in American Heart Journal that "so far no one has found any direct effect of nicotine on the heart other than that it somewhat increases heart out- put" (31). His position is supported by a German researcher, who stated in 1977 that "nicotine has been unjustifiably suspected for years" (44). The scientists who claim that nicotine causes heart disease have relied upon the results of animal experiments. But as one researcher who has conducted such experiments pointed out, daily dosages of nicotine used in these experi- ments were "equivalent to approximately 175 to 525 cigarettes per day in man; certainly an excessive and unrealistic amount" (45). In animal studies using realistic dosages, nicotine failed to initiate, exacerbate, or otherwise influence the atherogenic process in test animals (45, 46). Therefore, claims that nicotine, carbon monoxide or other components as found in tobacco smoke cause heart disease are not supported by the medical literature. As one British physician summarized the controversy, "Cigarette smoking is N 0 associated with a tendency to develop heart disease but there ''p 128

is no satisfactory epidemiological or experimental evidence to implicate the contents of smoke with disease of the coronary arteries" (47). ! Stroke HEW's 1975 report to Congress on smoking and health (48) said there has been "conflicting evidence on whether there is an increased risk of cerebrovascular disease due to smoking." In 1977, there was evidence from Johns Hopkins re- searchers that further confused the issue. Their epidemiological study of the differences in stroke mortality in three U.S. cities with low, intermediate and high rates of cerebrovascular disease was unable to account for them on the basis of smoking (49). They concluded that "there is a strong possibility" that physical and social environmental factors other than those presently known "may account for the observed geographic differences in mortality." The Johns Hopkins investigators noted that findings in past research work by others had also been "inconsistent". Although researchers in the past have studied risk factors associated with cardiovascular diseases as possible causes of stroke because they assumed all these diseases are related, this line of speculation has been discounted. "Cerebrovascular diseases are not simple extensions 129

Table 2 25-Year Trends of Age-Adjusted Mortality in the United States Cerebrovascular Diseases Heart Diseases Deaths/100,000 Decrease (%) Deaths/100,000 Decrease (%) 1950 88.8 -- 307.6 -- 1960 79.7 -10.2% 286.2 -6.8% 1969 68.3 -23.0% 262.3 -14.6% 1976 51.4 -42.4% 216.7 -29.4% Source: Tower 1978 (50) 130

or analogies of cardiovascular diseases," said one PHS official recently. "The fact is that strokes are neurological disorders and neuroepidemiological problems" (50). I I He is Donald B. Tower, who heads PHS's National Institute of Neurological and Communicative Disorders and Stroke, and the emphases were his, not ours. Dr. Tower addressed the Symposium on World Neurology in Montreal in September 1978. Stroke, he said, is "probably the most devastating and disabling of human disorders," preva- lent in every country regardless of economic, ethnic or cul- tural characteristics. And stroke "ought not to be equated epidemiologically or pathophysiologically" with cardiovascular disease. lished risk factor in countries like the U.S. and Japan, he said, pose "little, if any, risk for stroke" in the African countries. "When one examines stroke mortality over the past 25 years, the age-adjusted rate has decreased strikingly -- over 42 percent in the last quarter century [see Table 2] and a de- crease nearly 50 percent greater than that for heart disease." He said hypertension appears to be an important risk factor in stroke, but that there are countries like Nigeria and Senegal with high incidences of hypertension and of stroke but very low incidence of heart attacks. High cholesterol level, an estab- "We do not yet have the answers," said Dr. Tower. 131

Conclusion Since the work of the Framingham study helped produce the concept of "risk factors" supposedly associated with the development of CHD, it might be justified to conclude with a recent observation by its director and a PHS statistician who has also studied the data: A number of prominent cardiologists have lately expressed skepticism about the role of risk factors in cardiovascular disease and about the preventive and therapeutic efficacy of modifying them (16). The extent of current medical understanding of CHD causation was perhaps well described in the Annals of the New York Academy of Sciences: "The vast majority of individuals destined to develop and die from atherosclerotic disease do so for as yet unknown reasons "(51). 132

References for Cardiovascular Disease l. Dawber, T.R. The Framingham Study: A Need to Continue. Medical Tribune, Oct. 12, 1970. 2. Dawber, T.R., et al. Symposium on Arteriosclerosis: The Epidemiology of Coronary Heart Disease -- the Framingham Enquiry. Proceedings of the Royal Society of Medicine 55: 265-271, April 1962. 3. Kannel, W.B. Recent Findings of the Framingham Study. Resident and Staff Physician: 56-71, January 1978. 4. Werko, L. The Borderline Between Health and Disease, Prevention or Treatment? A Perspective. In: Waldenstrom et al, editors. Early Phases of Coronary Heart Disease: The Possibility of Prediction. Nordiska Bokhandelns Forlag, 1973, pp. 341-362. 5. Califano, J.A. Letter to all Members of Congress, Jan. 11, 1978. 6. Califano, J.A. Remarks of the Secretary of Health, Education and Welfare to the Association of State and Territorial Health Officials, Denver, April 20, 1978. 7. Califano, J.A. Remarks of the Secretary of Health, Educa- tion and Welfare before the Economic Club of Chicago, April 20, 1978. 8. Barnard, C. Heart Attack: You Don't Have to Die. Delacorte Press, New York, 1971. 9. Levy, R.I. Cardiovascular Disease: Past, Present, and Prospects for Prevention. Address to Mid-Atlantic Chapter of American Medical Writers Association, Bethesda, Md., Oct. 13, 1977, as reprinted in Medical Communications 6/2: Summer 1978. 10. Levy, R.I. Testimony before a subcommittee of the Commit- tee on Appropriations, U.S. House of Representatives, 95th Congress, Second Session, on Departments of Health, Educa- tion and Welfare appropriations for 1979, March 3, 1978. Part 3, pp. 317-444. 11. U.S. Public Health Service. Smoking and Health. Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington, U.S. Department of Health, Education and Welfare, PHS Pub. No. 1103, 1964, 387 pp. N Cn 0 ~ -p 4b- co N CIT N 133

12. Rossiter, A. Heart Disease. United Press International wire dispatch, 11:59AED, Oct. 24, 1978. 13. Stamler, J. Changes in CHD Risk Factors. Address to Con- ference on the Decline in Coronary Heart Disease Mortality, Division of Heart and Vascular Diseases, National Heart, Lung and Blood Institute, National Institutes of Health, U.S. Department of Health, Education and Welfare, Oct. 24-25, 1978. 14. Thom, T.J. Memorandum to Robert I. Levy, director, National Heart, Lung and Blood Institute, Oct. 27, 1977. 15. The Blue Sheet. Drug Research Reports 21/44: 4-5, Nov. 1, 1978. 16. Kannel, W.B., and Gordon, T. The Recent Decline in Fatal Heart Attacks: What Caused It? Primary Cardiology: 10-11, February 1978. 17. Forde, O.H., and Thelle, D.S. The Tromso Heart Study: Risk Factors for Coronary Heart Disease Related to the Occurrence of Myocardial Infarction in First Degree Relatives. American Journal of Epidemiology 105/3: 192-199, 1977. 18. Garcia-Palmieri, M.R., et al. Urban-Rural Differences in Coronary Heart Disease in a Low Incidence Area: The Puerto Rico Heart Study. American Journal of Epidemiology 107/3: 206-215, 1978. 19. Editorial: Virus Infection and Atherosclerosis. Lancet 2: 821-822, 1978. 20. Marmot, M.G., and Sy me, S.L. Acculturation and Coronary Heart Disease in Japanese-Americans. American Journal of Epidemiology 104/3: 225-247, Oct. 14, 1978. 21. Greenberg, J. The Americanization of Roseto. Science News 113/23: 378-379, 382, June 10, 1978. 22. Rosenman, R.H., and Friedman, M. Association of Specific Behavior Pattern in Women with Blood and Cardiovascular Findings. Circulation 24/5: 1173-1184, November 1961. 23. Matthews, K.A., et al. Competitive Drive, Pattern A, and Coronary Heart Disease: A Further Analysis of Some Data from the Western Collaborative Group Study. Journal of Chronic Diseases 30/8: 489-498, August 1977. 24. Blumenthal, J.A., et al. Type A Behavior Pattern and Coronary Atherosclerosis. Circulation 58/4: 634-639, N cn 0 ~ ~ ~ to N tfi (s? 134

October 1978. 25. Hartman, L.M. Behavioral Prevention of Ischemic Heart Disease. Canadian Medical Association Journal 119: 599-603, Sept. 23, 1978. 26. Frank, K.A., et al. Type A Behavior Pattern and Coronary Angiographic Findings. Journal of the American Medical Association 240/8: 761-763, Aug. 25, 1978. 27. Phillips, G.B. Relationship Between Serum Sex Hormones and Glucose, Insulin and Lipid Abnormalities in Men with Myocardial Infarction. Proceedings of the National Academy of Sciences USA 74/4: 1729-1733, April 1977. 28. Phillips, G.B. In: Shift in Balance of Sex Hormones Linked to Heart Attacks by Men, by Lawrence K. Altman. The New York Times, April 27, 1977. 29. Hormone Levels May be Involved in Heart Attacks. The New York Times, May 1, 1977. 30. Heller, R.F., and Jacobs, H.S. Coronary Heart Disease in Relation to Age, Sex, and the Menopause. British Medical Journal 1: 472-474, 1978. 31. Eskwith, I.S. Letter: American Heart Journal 90/6: 809-10, December 1975. 32. Horgan, J.H. Letter: Journal of the Irish Medical Association 69/13: 357, 361, Aug. 21, 1976. 33. Jenkins, C.D. Psychological and Social Precursors of Coronary Disease (First of Two Parts). New England Journal of Medicine 284/5: 244-255, Feb. 4, 1971. 34. Astrup, P., et al. Enhancing Influence of Carbon Mon- oxide on the Development of Atheromatosis in Cholesterol- Fed Rabbits. Journal of Atherosclerosis Research 7:. 343-354, 1967. 35. Wanstrup, J., et al. Acceleration of Spontaneous Intimal- Subintimal Changes in Rabbit Aorta by a Prolonged Moderate Carbon Monoxide Exposure. Acta Pathologica Microbiologica Scandanavica 75: 353-362, 1969. 36. Kj eldsen, K., et al. Ultrastructural Intimal Changes in the Rabbit Aorta after a Moderate Carbon Monoxide Exposure. Atherosclerosis 16/1: 67-82, 1972. fV CJ1 O ,.-. 37. Kjeldsen, K., et al. Effects of Carbon Monoxide on ~ Myocardium: Ultrastructural Changes in Rabbits After .p w N U1 4~b 135

Moderate, Chronic Exposure. Circulation Research 34/3: 339-348, March 1974. 38. Thomsen, H.K., and Kjeldsen, K. Threshold Limit for Carbon Monoxide-Induced Myocardial Damage. An Electron Microscopic Study in Rabbits. Archives of Environmental Health 30/12: 73-78, August 1974. 39. Thomsen, H.K., and Kjeldsen, K. Aortic Intimal Injury in Rabbits: An Evaluation of a Threshold Limit. Archives of Environmental Health 30/12: 604-607, 1975. 40. Wynder, E.L. Statement before the American Cancer Soci- ety's National Commission on Smoking and Public Policy, Philadelphia, June 16, 1977. 41. Stender, S., et al. The Effect of Carbon Monoxide on Cholesterol in the Aortic Wall of Rabbits. Atherosclerosis 28: 357-367, 1977. 42. Hugod, C., et al. The Influence of Carbon Monoxide on Intimal Morphology. Delivered at the International Conference on Atherosclerosis, Milan, Italy, Nov. 9, 1977. 43. Hugod, C., et al. Effect of Carbon Monoxide Exposure on Aortic and Coronary Intimal Morphology in the Rabbit: A Revaluation. Atherosclerosis 30: 333-342, August 1978. 44. Schievelbein, H. Bild Am Sonntag, March 13, 1977, p. 56. 45. Fisher, E.R., et al. Cigarette Smoking and Cholesterol: Atherosclerosis of Rabbits. Archives of Pathology 98: 418-421, December 1974. 46. Fisher, E.R., et al. Influence of Nicotine on Experimental Atherosclerosis and Its Determinants. Archives of Path- ology 96/5: 298-304, November 1973. 47. Gresham, G.A. Letter: The Cambridge (England) News, July 7, 1978. 48. U.S. Public Health Service. The Health Consequences of Smoking: 1975. U.S. Department of Health, Education and Welfare. Washington, DHEW Pub. No. (CDC) 76-8704, 1975, 235 pp. 49. Stolley, P.D., et al. Three-Area Epidemiological Study of Geographic Differences in Stroke Mortality. II. Results. N cn Stroke 8/5: 551-7, September/October 1977. 0 ~ ~ ~ w ~ cn En 136

50. Tower, D.B. Stroke: A Major Threat to World Health and to The Quality of Life. Presented at the Symposium on World Neurology: Tasks for the Future, Montreal, Que., Canada, Sept. 15, 1978. 51. Gutstein, W.H., et al. Neural Factors in Atherogenesis: Experimental Studies. of Sciences 275: 117-130, Annals of the New York Academy 1976. 137

---

---